]>
Commit | Line | Data |
---|---|---|
f3c0d7a5 A |
1 | // © 2016 and later: Unicode, Inc. and others. |
2 | // License & terms of use: http://www.unicode.org/copyright.html | |
4388f060 A |
3 | /* |
4 | ******************************************************************************* | |
b331163b | 5 | * Copyright (C) 2011-2015, International Business Machines Corporation and |
57a6839d | 6 | * others. All Rights Reserved. |
4388f060 A |
7 | ******************************************************************************* |
8 | */ | |
9 | ||
10 | #include "unicode/utypes.h" | |
11 | ||
12 | #if !UCONFIG_NO_FORMATTING | |
13 | ||
14 | #include "unicode/calendar.h" | |
15 | #include "unicode/tzfmt.h" | |
16 | #include "unicode/numsys.h" | |
f3c0d7a5 | 17 | #include "unicode/strenum.h" |
4388f060 A |
18 | #include "unicode/uchar.h" |
19 | #include "unicode/udat.h" | |
f3c0d7a5 | 20 | #include "unicode/ustring.h" |
0f5d89e8 | 21 | #include "unicode/utf16.h" |
4388f060 A |
22 | #include "tzgnames.h" |
23 | #include "cmemory.h" | |
24 | #include "cstring.h" | |
25 | #include "putilimp.h" | |
26 | #include "uassert.h" | |
27 | #include "ucln_in.h" | |
28 | #include "umutex.h" | |
29 | #include "uresimp.h" | |
30 | #include "ureslocs.h" | |
31 | #include "uvector.h" | |
32 | #include "zonemeta.h" | |
51004dcb | 33 | #include "tznames_impl.h" // TextTrieMap |
0f5d89e8 | 34 | #include "patternprops.h" |
4388f060 A |
35 | |
36 | U_NAMESPACE_BEGIN | |
37 | ||
51004dcb A |
38 | // Bit flags used by the parse method. |
39 | // The order must match UTimeZoneFormatStyle enum. | |
40 | #define ISO_Z_STYLE_FLAG 0x0080 | |
41 | #define ISO_LOCAL_STYLE_FLAG 0x0100 | |
42 | static const int16_t STYLE_PARSE_FLAGS[] = { | |
43 | 0x0001, // UTZFMT_STYLE_GENERIC_LOCATION, | |
44 | 0x0002, // UTZFMT_STYLE_GENERIC_LONG, | |
45 | 0x0004, // UTZFMT_STYLE_GENERIC_SHORT, | |
46 | 0x0008, // UTZFMT_STYLE_SPECIFIC_LONG, | |
47 | 0x0010, // UTZFMT_STYLE_SPECIFIC_SHORT, | |
48 | 0x0020, // UTZFMT_STYLE_LOCALIZED_GMT, | |
49 | 0x0040, // UTZFMT_STYLE_LOCALIZED_GMT_SHORT, | |
50 | ISO_Z_STYLE_FLAG, // UTZFMT_STYLE_ISO_BASIC_SHORT, | |
51 | ISO_LOCAL_STYLE_FLAG, // UTZFMT_STYLE_ISO_BASIC_LOCAL_SHORT, | |
52 | ISO_Z_STYLE_FLAG, // UTZFMT_STYLE_ISO_BASIC_FIXED, | |
53 | ISO_LOCAL_STYLE_FLAG, // UTZFMT_STYLE_ISO_BASIC_LOCAL_FIXED, | |
54 | ISO_Z_STYLE_FLAG, // UTZFMT_STYLE_ISO_BASIC_FULL, | |
55 | ISO_LOCAL_STYLE_FLAG, // UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL, | |
56 | ISO_Z_STYLE_FLAG, // UTZFMT_STYLE_ISO_EXTENDED_FIXED, | |
57 | ISO_LOCAL_STYLE_FLAG, // UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FIXED, | |
58 | ISO_Z_STYLE_FLAG, // UTZFMT_STYLE_ISO_EXTENDED_FULL, | |
59 | ISO_LOCAL_STYLE_FLAG, // UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FULL, | |
60 | 0x0200, // UTZFMT_STYLE_ZONE_ID, | |
61 | 0x0400, // UTZFMT_STYLE_ZONE_ID_SHORT, | |
62 | 0x0800 // UTZFMT_STYLE_EXEMPLAR_LOCATION | |
63 | }; | |
64 | ||
4388f060 A |
65 | static const char gZoneStringsTag[] = "zoneStrings"; |
66 | static const char gGmtFormatTag[]= "gmtFormat"; | |
67 | static const char gGmtZeroFormatTag[] = "gmtZeroFormat"; | |
68 | static const char gHourFormatTag[]= "hourFormat"; | |
69 | ||
70 | static const UChar TZID_GMT[] = {0x0045, 0x0074, 0x0063, 0x002F, 0x0047, 0x004D, 0x0054, 0}; // Etc/GMT | |
51004dcb A |
71 | static const UChar UNKNOWN_ZONE_ID[] = { |
72 | 0x0045, 0x0074, 0x0063, 0x002F, 0x0055, 0x006E, 0x006B, 0x006E, 0x006F, 0x0077, 0x006E, 0}; // Etc/Unknown | |
73 | static const UChar UNKNOWN_SHORT_ZONE_ID[] = {0x0075, 0x006E, 0x006B, 0}; // unk | |
74 | static const UChar UNKNOWN_LOCATION[] = {0x0055, 0x006E, 0x006B, 0x006E, 0x006F, 0x0077, 0x006E, 0}; // Unknown | |
4388f060 A |
75 | |
76 | static const UChar DEFAULT_GMT_PATTERN[] = {0x0047, 0x004D, 0x0054, 0x007B, 0x0030, 0x007D, 0}; // GMT{0} | |
51004dcb A |
77 | //static const UChar DEFAULT_GMT_ZERO[] = {0x0047, 0x004D, 0x0054, 0}; // GMT |
78 | static const UChar DEFAULT_GMT_POSITIVE_HM[] = {0x002B, 0x0048, 0x003A, 0x006D, 0x006D, 0}; // +H:mm | |
79 | static const UChar DEFAULT_GMT_POSITIVE_HMS[] = {0x002B, 0x0048, 0x003A, 0x006D, 0x006D, 0x003A, 0x0073, 0x0073, 0}; // +H:mm:ss | |
80 | static const UChar DEFAULT_GMT_NEGATIVE_HM[] = {0x002D, 0x0048, 0x003A, 0x006D, 0x006D, 0}; // -H:mm | |
81 | static const UChar DEFAULT_GMT_NEGATIVE_HMS[] = {0x002D, 0x0048, 0x003A, 0x006D, 0x006D, 0x003A, 0x0073, 0x0073, 0}; // -H:mm:ss | |
82 | static const UChar DEFAULT_GMT_POSITIVE_H[] = {0x002B, 0x0048, 0}; // +H | |
83 | static const UChar DEFAULT_GMT_NEGATIVE_H[] = {0x002D, 0x0048, 0}; // -H | |
4388f060 A |
84 | |
85 | static const UChar32 DEFAULT_GMT_DIGITS[] = { | |
86 | 0x0030, 0x0031, 0x0032, 0x0033, 0x0034, | |
87 | 0x0035, 0x0036, 0x0037, 0x0038, 0x0039 | |
88 | }; | |
89 | ||
90 | static const UChar DEFAULT_GMT_OFFSET_SEP = 0x003A; // ':' | |
91 | ||
92 | static const UChar ARG0[] = {0x007B, 0x0030, 0x007D}; // "{0}" | |
51004dcb | 93 | static const int32_t ARG0_LEN = 3; |
4388f060 A |
94 | |
95 | static const UChar DEFAULT_GMT_OFFSET_MINUTE_PATTERN[] = {0x006D, 0x006D, 0}; // "mm" | |
96 | static const UChar DEFAULT_GMT_OFFSET_SECOND_PATTERN[] = {0x0073, 0x0073, 0}; // "ss" | |
97 | ||
98 | static const UChar ALT_GMT_STRINGS[][4] = { | |
99 | {0x0047, 0x004D, 0x0054, 0}, // GMT | |
100 | {0x0055, 0x0054, 0x0043, 0}, // UTC | |
101 | {0x0055, 0x0054, 0, 0}, // UT | |
102 | {0, 0, 0, 0} | |
103 | }; | |
104 | ||
105 | // Order of GMT offset pattern parsing, *_HMS must be evaluated first | |
106 | // because *_HM is most likely a substring of *_HMS | |
107 | static const int32_t PARSE_GMT_OFFSET_TYPES[] = { | |
108 | UTZFMT_PAT_POSITIVE_HMS, | |
109 | UTZFMT_PAT_NEGATIVE_HMS, | |
110 | UTZFMT_PAT_POSITIVE_HM, | |
111 | UTZFMT_PAT_NEGATIVE_HM, | |
51004dcb A |
112 | UTZFMT_PAT_POSITIVE_H, |
113 | UTZFMT_PAT_NEGATIVE_H, | |
4388f060 A |
114 | -1 |
115 | }; | |
116 | ||
117 | static const UChar SINGLEQUOTE = 0x0027; | |
118 | static const UChar PLUS = 0x002B; | |
119 | static const UChar MINUS = 0x002D; | |
120 | static const UChar ISO8601_UTC = 0x005A; // 'Z' | |
121 | static const UChar ISO8601_SEP = 0x003A; // ':' | |
122 | ||
123 | static const int32_t MILLIS_PER_HOUR = 60 * 60 * 1000; | |
124 | static const int32_t MILLIS_PER_MINUTE = 60 * 1000; | |
125 | static const int32_t MILLIS_PER_SECOND = 1000; | |
126 | ||
127 | // Maximum offset (exclusive) in millisecond supported by offset formats | |
128 | static int32_t MAX_OFFSET = 24 * MILLIS_PER_HOUR; | |
129 | ||
130 | // Maximum values for GMT offset fields | |
131 | static const int32_t MAX_OFFSET_HOUR = 23; | |
132 | static const int32_t MAX_OFFSET_MINUTE = 59; | |
133 | static const int32_t MAX_OFFSET_SECOND = 59; | |
134 | ||
135 | static const int32_t UNKNOWN_OFFSET = 0x7FFFFFFF; | |
136 | ||
51004dcb | 137 | static const int32_t ALL_SIMPLE_NAME_TYPES = UTZNM_LONG_STANDARD | UTZNM_LONG_DAYLIGHT | UTZNM_SHORT_STANDARD | UTZNM_SHORT_DAYLIGHT | UTZNM_EXEMPLAR_LOCATION; |
4388f060 A |
138 | static const int32_t ALL_GENERIC_NAME_TYPES = UTZGNM_LOCATION | UTZGNM_LONG | UTZGNM_SHORT; |
139 | ||
4388f060 A |
140 | #define DIGIT_VAL(c) (0x0030 <= (c) && (c) <= 0x0039 ? (c) - 0x0030 : -1) |
141 | #define MAX_OFFSET_DIGITS 6 | |
142 | ||
51004dcb A |
143 | // Time Zone ID/Short ID trie |
144 | static TextTrieMap *gZoneIdTrie = NULL; | |
57a6839d | 145 | static icu::UInitOnce gZoneIdTrieInitOnce = U_INITONCE_INITIALIZER; |
51004dcb A |
146 | |
147 | static TextTrieMap *gShortZoneIdTrie = NULL; | |
57a6839d | 148 | static icu::UInitOnce gShortZoneIdTrieInitOnce = U_INITONCE_INITIALIZER; |
51004dcb | 149 | |
3d1f044b A |
150 | static UMutex *gLock() { |
151 | static UMutex *m = STATIC_NEW(UMutex); | |
152 | return m; | |
153 | } | |
51004dcb A |
154 | |
155 | U_CDECL_BEGIN | |
156 | /** | |
157 | * Cleanup callback func | |
158 | */ | |
159 | static UBool U_CALLCONV tzfmt_cleanup(void) | |
160 | { | |
161 | if (gZoneIdTrie != NULL) { | |
162 | delete gZoneIdTrie; | |
163 | } | |
164 | gZoneIdTrie = NULL; | |
57a6839d | 165 | gZoneIdTrieInitOnce.reset(); |
51004dcb A |
166 | |
167 | if (gShortZoneIdTrie != NULL) { | |
168 | delete gShortZoneIdTrie; | |
169 | } | |
170 | gShortZoneIdTrie = NULL; | |
57a6839d | 171 | gShortZoneIdTrieInitOnce.reset(); |
51004dcb A |
172 | |
173 | return TRUE; | |
174 | } | |
175 | U_CDECL_END | |
4388f060 A |
176 | |
177 | // ------------------------------------------------------------------ | |
178 | // GMTOffsetField | |
179 | // | |
180 | // This class represents a localized GMT offset pattern | |
181 | // item and used by TimeZoneFormat | |
182 | // ------------------------------------------------------------------ | |
183 | class GMTOffsetField : public UMemory { | |
184 | public: | |
185 | enum FieldType { | |
186 | TEXT = 0, | |
187 | HOUR = 1, | |
188 | MINUTE = 2, | |
189 | SECOND = 4 | |
190 | }; | |
191 | ||
192 | virtual ~GMTOffsetField(); | |
193 | ||
194 | static GMTOffsetField* createText(const UnicodeString& text, UErrorCode& status); | |
195 | static GMTOffsetField* createTimeField(FieldType type, uint8_t width, UErrorCode& status); | |
196 | static UBool isValid(FieldType type, int32_t width); | |
197 | static FieldType getTypeByLetter(UChar ch); | |
198 | ||
199 | FieldType getType() const; | |
200 | uint8_t getWidth() const; | |
201 | const UChar* getPatternText(void) const; | |
202 | ||
203 | private: | |
204 | UChar* fText; | |
205 | FieldType fType; | |
206 | uint8_t fWidth; | |
207 | ||
208 | GMTOffsetField(); | |
209 | }; | |
210 | ||
211 | GMTOffsetField::GMTOffsetField() | |
212 | : fText(NULL), fType(TEXT), fWidth(0) { | |
213 | } | |
214 | ||
215 | GMTOffsetField::~GMTOffsetField() { | |
216 | if (fText) { | |
217 | uprv_free(fText); | |
218 | } | |
219 | } | |
220 | ||
221 | GMTOffsetField* | |
222 | GMTOffsetField::createText(const UnicodeString& text, UErrorCode& status) { | |
223 | if (U_FAILURE(status)) { | |
224 | return NULL; | |
225 | } | |
226 | GMTOffsetField* result = new GMTOffsetField(); | |
227 | if (result == NULL) { | |
228 | status = U_MEMORY_ALLOCATION_ERROR; | |
229 | return NULL; | |
230 | } | |
231 | ||
232 | int32_t len = text.length(); | |
233 | result->fText = (UChar*)uprv_malloc((len + 1) * sizeof(UChar)); | |
234 | if (result->fText == NULL) { | |
235 | status = U_MEMORY_ALLOCATION_ERROR; | |
236 | delete result; | |
237 | return NULL; | |
238 | } | |
239 | u_strncpy(result->fText, text.getBuffer(), len); | |
240 | result->fText[len] = 0; | |
241 | result->fType = TEXT; | |
242 | ||
243 | return result; | |
244 | } | |
245 | ||
246 | GMTOffsetField* | |
247 | GMTOffsetField::createTimeField(FieldType type, uint8_t width, UErrorCode& status) { | |
248 | U_ASSERT(type != TEXT); | |
249 | if (U_FAILURE(status)) { | |
250 | return NULL; | |
251 | } | |
252 | GMTOffsetField* result = new GMTOffsetField(); | |
253 | if (result == NULL) { | |
254 | status = U_MEMORY_ALLOCATION_ERROR; | |
255 | return NULL; | |
256 | } | |
257 | ||
258 | result->fType = type; | |
259 | result->fWidth = width; | |
260 | ||
261 | return result; | |
262 | } | |
263 | ||
264 | UBool | |
265 | GMTOffsetField::isValid(FieldType type, int32_t width) { | |
266 | switch (type) { | |
267 | case HOUR: | |
268 | return (width == 1 || width == 2); | |
269 | case MINUTE: | |
270 | case SECOND: | |
271 | return (width == 2); | |
272 | default: | |
3d1f044b | 273 | UPRV_UNREACHABLE; |
4388f060 A |
274 | } |
275 | return (width > 0); | |
276 | } | |
277 | ||
278 | GMTOffsetField::FieldType | |
279 | GMTOffsetField::getTypeByLetter(UChar ch) { | |
280 | if (ch == 0x0048 /* H */) { | |
281 | return HOUR; | |
282 | } else if (ch == 0x006D /* m */) { | |
283 | return MINUTE; | |
284 | } else if (ch == 0x0073 /* s */) { | |
285 | return SECOND; | |
286 | } | |
287 | return TEXT; | |
288 | } | |
289 | ||
290 | inline GMTOffsetField::FieldType | |
291 | GMTOffsetField::getType() const { | |
292 | return fType; | |
293 | } | |
294 | ||
295 | inline uint8_t | |
296 | GMTOffsetField::getWidth() const { | |
297 | return fWidth; | |
298 | } | |
299 | ||
300 | inline const UChar* | |
301 | GMTOffsetField::getPatternText(void) const { | |
302 | return fText; | |
303 | } | |
304 | ||
305 | ||
306 | U_CDECL_BEGIN | |
307 | static void U_CALLCONV | |
308 | deleteGMTOffsetField(void *obj) { | |
309 | delete static_cast<GMTOffsetField *>(obj); | |
310 | } | |
311 | U_CDECL_END | |
312 | ||
313 | ||
314 | // ------------------------------------------------------------------ | |
315 | // TimeZoneFormat | |
316 | // ------------------------------------------------------------------ | |
317 | UOBJECT_DEFINE_RTTI_IMPLEMENTATION(TimeZoneFormat) | |
318 | ||
319 | TimeZoneFormat::TimeZoneFormat(const Locale& locale, UErrorCode& status) | |
b331163b A |
320 | : fLocale(locale), fTimeZoneNames(NULL), fTimeZoneGenericNames(NULL), |
321 | fDefParseOptionFlags(0), fTZDBTimeZoneNames(NULL) { | |
4388f060 | 322 | |
51004dcb | 323 | for (int32_t i = 0; i < UTZFMT_PAT_COUNT; i++) { |
4388f060 A |
324 | fGMTOffsetPatternItems[i] = NULL; |
325 | } | |
326 | ||
327 | const char* region = fLocale.getCountry(); | |
0f5d89e8 | 328 | int32_t regionLen = static_cast<int32_t>(uprv_strlen(region)); |
4388f060 A |
329 | if (regionLen == 0) { |
330 | char loc[ULOC_FULLNAME_CAPACITY]; | |
331 | uloc_addLikelySubtags(fLocale.getName(), loc, sizeof(loc), &status); | |
332 | ||
333 | regionLen = uloc_getCountry(loc, fTargetRegion, sizeof(fTargetRegion), &status); | |
334 | if (U_SUCCESS(status)) { | |
335 | fTargetRegion[regionLen] = 0; | |
336 | } else { | |
337 | return; | |
338 | } | |
339 | } else if (regionLen < (int32_t)sizeof(fTargetRegion)) { | |
340 | uprv_strcpy(fTargetRegion, region); | |
341 | } else { | |
342 | fTargetRegion[0] = 0; | |
343 | } | |
344 | ||
345 | fTimeZoneNames = TimeZoneNames::createInstance(locale, status); | |
346 | // fTimeZoneGenericNames is lazily instantiated | |
51004dcb A |
347 | if (U_FAILURE(status)) { |
348 | return; | |
349 | } | |
4388f060 A |
350 | |
351 | const UChar* gmtPattern = NULL; | |
352 | const UChar* hourFormats = NULL; | |
353 | ||
354 | UResourceBundle *zoneBundle = ures_open(U_ICUDATA_ZONE, locale.getName(), &status); | |
355 | UResourceBundle *zoneStringsArray = ures_getByKeyWithFallback(zoneBundle, gZoneStringsTag, NULL, &status); | |
356 | if (U_SUCCESS(status)) { | |
357 | const UChar* resStr; | |
358 | int32_t len; | |
359 | resStr = ures_getStringByKeyWithFallback(zoneStringsArray, gGmtFormatTag, &len, &status); | |
360 | if (len > 0) { | |
361 | gmtPattern = resStr; | |
362 | } | |
363 | resStr = ures_getStringByKeyWithFallback(zoneStringsArray, gGmtZeroFormatTag, &len, &status); | |
364 | if (len > 0) { | |
365 | fGMTZeroFormat.setTo(TRUE, resStr, len); | |
366 | } | |
367 | resStr = ures_getStringByKeyWithFallback(zoneStringsArray, gHourFormatTag, &len, &status); | |
368 | if (len > 0) { | |
369 | hourFormats = resStr; | |
370 | } | |
371 | ures_close(zoneStringsArray); | |
372 | ures_close(zoneBundle); | |
373 | } | |
374 | ||
375 | if (gmtPattern == NULL) { | |
376 | gmtPattern = DEFAULT_GMT_PATTERN; | |
377 | } | |
b331163b | 378 | initGMTPattern(UnicodeString(TRUE, gmtPattern, -1), status); |
4388f060 | 379 | |
51004dcb | 380 | UBool useDefaultOffsetPatterns = TRUE; |
4388f060 A |
381 | if (hourFormats) { |
382 | UChar *sep = u_strchr(hourFormats, (UChar)0x003B /* ';' */); | |
383 | if (sep != NULL) { | |
51004dcb | 384 | UErrorCode tmpStatus = U_ZERO_ERROR; |
4388f060 A |
385 | fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_HM].setTo(FALSE, hourFormats, (int32_t)(sep - hourFormats)); |
386 | fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_HM].setTo(TRUE, sep + 1, -1); | |
51004dcb A |
387 | expandOffsetPattern(fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_HM], fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_HMS], tmpStatus); |
388 | expandOffsetPattern(fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_HM], fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_HMS], tmpStatus); | |
389 | truncateOffsetPattern(fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_HM], fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_H], tmpStatus); | |
390 | truncateOffsetPattern(fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_HM], fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_H], tmpStatus); | |
391 | if (U_SUCCESS(tmpStatus)) { | |
392 | useDefaultOffsetPatterns = FALSE; | |
393 | } | |
4388f060 A |
394 | } |
395 | } | |
51004dcb A |
396 | if (useDefaultOffsetPatterns) { |
397 | fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_H].setTo(TRUE, DEFAULT_GMT_POSITIVE_H, -1); | |
4388f060 A |
398 | fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_HM].setTo(TRUE, DEFAULT_GMT_POSITIVE_HM, -1); |
399 | fGMTOffsetPatterns[UTZFMT_PAT_POSITIVE_HMS].setTo(TRUE, DEFAULT_GMT_POSITIVE_HMS, -1); | |
51004dcb | 400 | fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_H].setTo(TRUE, DEFAULT_GMT_NEGATIVE_H, -1); |
4388f060 A |
401 | fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_HM].setTo(TRUE, DEFAULT_GMT_NEGATIVE_HM, -1); |
402 | fGMTOffsetPatterns[UTZFMT_PAT_NEGATIVE_HMS].setTo(TRUE, DEFAULT_GMT_NEGATIVE_HMS, -1); | |
403 | } | |
404 | initGMTOffsetPatterns(status); | |
405 | ||
406 | NumberingSystem* ns = NumberingSystem::createInstance(locale, status); | |
407 | UBool useDefDigits = TRUE; | |
408 | if (ns && !ns->isAlgorithmic()) { | |
409 | UnicodeString digits = ns->getDescription(); | |
410 | useDefDigits = !toCodePoints(digits, fGMTOffsetDigits, 10); | |
411 | } | |
412 | if (useDefDigits) { | |
413 | uprv_memcpy(fGMTOffsetDigits, DEFAULT_GMT_DIGITS, sizeof(UChar32) * 10); | |
414 | } | |
415 | delete ns; | |
416 | } | |
417 | ||
418 | TimeZoneFormat::TimeZoneFormat(const TimeZoneFormat& other) | |
b331163b A |
419 | : Format(other), fTimeZoneNames(NULL), fTimeZoneGenericNames(NULL), |
420 | fTZDBTimeZoneNames(NULL) { | |
4388f060 | 421 | |
51004dcb | 422 | for (int32_t i = 0; i < UTZFMT_PAT_COUNT; i++) { |
4388f060 A |
423 | fGMTOffsetPatternItems[i] = NULL; |
424 | } | |
425 | *this = other; | |
426 | } | |
427 | ||
428 | ||
429 | TimeZoneFormat::~TimeZoneFormat() { | |
430 | delete fTimeZoneNames; | |
431 | delete fTimeZoneGenericNames; | |
b331163b | 432 | delete fTZDBTimeZoneNames; |
51004dcb | 433 | for (int32_t i = 0; i < UTZFMT_PAT_COUNT; i++) { |
4388f060 A |
434 | delete fGMTOffsetPatternItems[i]; |
435 | } | |
4388f060 A |
436 | } |
437 | ||
438 | TimeZoneFormat& | |
439 | TimeZoneFormat::operator=(const TimeZoneFormat& other) { | |
440 | if (this == &other) { | |
441 | return *this; | |
442 | } | |
443 | ||
444 | delete fTimeZoneNames; | |
445 | delete fTimeZoneGenericNames; | |
446 | fTimeZoneGenericNames = NULL; | |
b331163b A |
447 | delete fTZDBTimeZoneNames; |
448 | fTZDBTimeZoneNames = NULL; | |
4388f060 A |
449 | |
450 | fLocale = other.fLocale; | |
451 | uprv_memcpy(fTargetRegion, other.fTargetRegion, sizeof(fTargetRegion)); | |
452 | ||
453 | fTimeZoneNames = other.fTimeZoneNames->clone(); | |
454 | if (other.fTimeZoneGenericNames) { | |
57a6839d | 455 | // TODO: this test has dubious thread safety. |
4388f060 A |
456 | fTimeZoneGenericNames = other.fTimeZoneGenericNames->clone(); |
457 | } | |
458 | ||
459 | fGMTPattern = other.fGMTPattern; | |
460 | fGMTPatternPrefix = other.fGMTPatternPrefix; | |
461 | fGMTPatternSuffix = other.fGMTPatternSuffix; | |
462 | ||
463 | UErrorCode status = U_ZERO_ERROR; | |
51004dcb | 464 | for (int32_t i = 0; i < UTZFMT_PAT_COUNT; i++) { |
4388f060 A |
465 | fGMTOffsetPatterns[i] = other.fGMTOffsetPatterns[i]; |
466 | delete fGMTOffsetPatternItems[i]; | |
b331163b | 467 | fGMTOffsetPatternItems[i] = NULL; |
4388f060 A |
468 | } |
469 | initGMTOffsetPatterns(status); | |
470 | U_ASSERT(U_SUCCESS(status)); | |
471 | ||
472 | fGMTZeroFormat = other.fGMTZeroFormat; | |
473 | ||
474 | uprv_memcpy(fGMTOffsetDigits, other.fGMTOffsetDigits, sizeof(fGMTOffsetDigits)); | |
475 | ||
476 | fDefParseOptionFlags = other.fDefParseOptionFlags; | |
477 | ||
478 | return *this; | |
479 | } | |
480 | ||
481 | ||
482 | UBool | |
483 | TimeZoneFormat::operator==(const Format& other) const { | |
484 | TimeZoneFormat* tzfmt = (TimeZoneFormat*)&other; | |
485 | ||
486 | UBool isEqual = | |
487 | fLocale == tzfmt->fLocale | |
488 | && fGMTPattern == tzfmt->fGMTPattern | |
489 | && fGMTZeroFormat == tzfmt->fGMTZeroFormat | |
490 | && *fTimeZoneNames == *tzfmt->fTimeZoneNames; | |
491 | ||
51004dcb | 492 | for (int32_t i = 0; i < UTZFMT_PAT_COUNT && isEqual; i++) { |
4388f060 A |
493 | isEqual = fGMTOffsetPatterns[i] == tzfmt->fGMTOffsetPatterns[i]; |
494 | } | |
495 | for (int32_t i = 0; i < 10 && isEqual; i++) { | |
496 | isEqual = fGMTOffsetDigits[i] == tzfmt->fGMTOffsetDigits[i]; | |
497 | } | |
498 | // TODO | |
499 | // Check fTimeZoneGenericNames. For now, | |
500 | // if fTimeZoneNames is same, fTimeZoneGenericNames should | |
501 | // be also equivalent. | |
502 | return isEqual; | |
503 | } | |
504 | ||
505 | Format* | |
506 | TimeZoneFormat::clone() const { | |
507 | return new TimeZoneFormat(*this); | |
508 | } | |
509 | ||
510 | TimeZoneFormat* U_EXPORT2 | |
511 | TimeZoneFormat::createInstance(const Locale& locale, UErrorCode& status) { | |
512 | TimeZoneFormat* tzfmt = new TimeZoneFormat(locale, status); | |
513 | if (U_SUCCESS(status)) { | |
514 | return tzfmt; | |
515 | } | |
516 | delete tzfmt; | |
517 | return NULL; | |
518 | } | |
519 | ||
520 | // ------------------------------------------------------------------ | |
521 | // Setter and Getter | |
522 | ||
523 | const TimeZoneNames* | |
524 | TimeZoneFormat::getTimeZoneNames() const { | |
525 | return (const TimeZoneNames*)fTimeZoneNames; | |
526 | } | |
527 | ||
528 | void | |
529 | TimeZoneFormat::adoptTimeZoneNames(TimeZoneNames *tznames) { | |
530 | delete fTimeZoneNames; | |
531 | fTimeZoneNames = tznames; | |
532 | ||
533 | // TODO - We should also update fTimeZoneGenericNames | |
534 | } | |
535 | ||
536 | void | |
537 | TimeZoneFormat::setTimeZoneNames(const TimeZoneNames &tznames) { | |
538 | delete fTimeZoneNames; | |
539 | fTimeZoneNames = tznames.clone(); | |
540 | ||
541 | // TODO - We should also update fTimeZoneGenericNames | |
542 | } | |
543 | ||
544 | void | |
51004dcb | 545 | TimeZoneFormat::setDefaultParseOptions(uint32_t flags) { |
4388f060 A |
546 | fDefParseOptionFlags = flags; |
547 | } | |
548 | ||
51004dcb | 549 | uint32_t |
4388f060 A |
550 | TimeZoneFormat::getDefaultParseOptions(void) const { |
551 | return fDefParseOptionFlags; | |
552 | } | |
553 | ||
554 | ||
555 | UnicodeString& | |
556 | TimeZoneFormat::getGMTPattern(UnicodeString& pattern) const { | |
557 | return pattern.setTo(fGMTPattern); | |
558 | } | |
559 | ||
560 | void | |
561 | TimeZoneFormat::setGMTPattern(const UnicodeString& pattern, UErrorCode& status) { | |
562 | initGMTPattern(pattern, status); | |
563 | } | |
564 | ||
565 | UnicodeString& | |
566 | TimeZoneFormat::getGMTOffsetPattern(UTimeZoneFormatGMTOffsetPatternType type, UnicodeString& pattern) const { | |
567 | return pattern.setTo(fGMTOffsetPatterns[type]); | |
568 | } | |
569 | ||
570 | void | |
571 | TimeZoneFormat::setGMTOffsetPattern(UTimeZoneFormatGMTOffsetPatternType type, const UnicodeString& pattern, UErrorCode& status) { | |
572 | if (U_FAILURE(status)) { | |
573 | return; | |
574 | } | |
575 | if (pattern == fGMTOffsetPatterns[type]) { | |
576 | // No need to reset | |
577 | return; | |
578 | } | |
579 | ||
51004dcb A |
580 | OffsetFields required = FIELDS_HM; |
581 | switch (type) { | |
582 | case UTZFMT_PAT_POSITIVE_H: | |
583 | case UTZFMT_PAT_NEGATIVE_H: | |
584 | required = FIELDS_H; | |
585 | break; | |
586 | case UTZFMT_PAT_POSITIVE_HM: | |
587 | case UTZFMT_PAT_NEGATIVE_HM: | |
588 | required = FIELDS_HM; | |
589 | break; | |
590 | case UTZFMT_PAT_POSITIVE_HMS: | |
591 | case UTZFMT_PAT_NEGATIVE_HMS: | |
592 | required = FIELDS_HMS; | |
593 | break; | |
594 | default: | |
3d1f044b | 595 | UPRV_UNREACHABLE; |
51004dcb | 596 | } |
4388f060 A |
597 | |
598 | UVector* patternItems = parseOffsetPattern(pattern, required, status); | |
599 | if (patternItems == NULL) { | |
600 | return; | |
601 | } | |
602 | ||
603 | fGMTOffsetPatterns[type].setTo(pattern); | |
604 | delete fGMTOffsetPatternItems[type]; | |
605 | fGMTOffsetPatternItems[type] = patternItems; | |
51004dcb | 606 | checkAbuttingHoursAndMinutes(); |
4388f060 A |
607 | } |
608 | ||
609 | UnicodeString& | |
610 | TimeZoneFormat::getGMTOffsetDigits(UnicodeString& digits) const { | |
611 | digits.remove(); | |
612 | for (int32_t i = 0; i < 10; i++) { | |
613 | digits.append(fGMTOffsetDigits[i]); | |
614 | } | |
615 | return digits; | |
616 | } | |
617 | ||
618 | void | |
619 | TimeZoneFormat::setGMTOffsetDigits(const UnicodeString& digits, UErrorCode& status) { | |
620 | if (U_FAILURE(status)) { | |
621 | return; | |
622 | } | |
623 | UChar32 digitArray[10]; | |
624 | if (!toCodePoints(digits, digitArray, 10)) { | |
625 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
626 | return; | |
627 | } | |
628 | uprv_memcpy(fGMTOffsetDigits, digitArray, sizeof(UChar32)*10); | |
629 | } | |
630 | ||
631 | UnicodeString& | |
632 | TimeZoneFormat::getGMTZeroFormat(UnicodeString& gmtZeroFormat) const { | |
633 | return gmtZeroFormat.setTo(fGMTZeroFormat); | |
634 | } | |
635 | ||
636 | void | |
637 | TimeZoneFormat::setGMTZeroFormat(const UnicodeString& gmtZeroFormat, UErrorCode& status) { | |
638 | if (U_SUCCESS(status)) { | |
639 | if (gmtZeroFormat.isEmpty()) { | |
640 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
641 | } else if (gmtZeroFormat != fGMTZeroFormat) { | |
642 | fGMTZeroFormat.setTo(gmtZeroFormat); | |
643 | } | |
644 | } | |
645 | } | |
646 | ||
647 | // ------------------------------------------------------------------ | |
648 | // Format and Parse | |
649 | ||
650 | UnicodeString& | |
651 | TimeZoneFormat::format(UTimeZoneFormatStyle style, const TimeZone& tz, UDate date, | |
652 | UnicodeString& name, UTimeZoneFormatTimeType* timeType /* = NULL */) const { | |
653 | if (timeType) { | |
654 | *timeType = UTZFMT_TIME_TYPE_UNKNOWN; | |
655 | } | |
57a6839d A |
656 | |
657 | UBool noOffsetFormatFallback = FALSE; | |
658 | ||
4388f060 A |
659 | switch (style) { |
660 | case UTZFMT_STYLE_GENERIC_LOCATION: | |
661 | formatGeneric(tz, UTZGNM_LOCATION, date, name); | |
662 | break; | |
663 | case UTZFMT_STYLE_GENERIC_LONG: | |
664 | formatGeneric(tz, UTZGNM_LONG, date, name); | |
665 | break; | |
666 | case UTZFMT_STYLE_GENERIC_SHORT: | |
667 | formatGeneric(tz, UTZGNM_SHORT, date, name); | |
668 | break; | |
669 | case UTZFMT_STYLE_SPECIFIC_LONG: | |
670 | formatSpecific(tz, UTZNM_LONG_STANDARD, UTZNM_LONG_DAYLIGHT, date, name, timeType); | |
671 | break; | |
672 | case UTZFMT_STYLE_SPECIFIC_SHORT: | |
673 | formatSpecific(tz, UTZNM_SHORT_STANDARD, UTZNM_SHORT_DAYLIGHT, date, name, timeType); | |
674 | break; | |
57a6839d A |
675 | |
676 | case UTZFMT_STYLE_ZONE_ID: | |
677 | tz.getID(name); | |
678 | noOffsetFormatFallback = TRUE; | |
679 | break; | |
680 | case UTZFMT_STYLE_ZONE_ID_SHORT: | |
681 | { | |
682 | const UChar* shortID = ZoneMeta::getShortID(tz); | |
683 | if (shortID == NULL) { | |
684 | shortID = UNKNOWN_SHORT_ZONE_ID; | |
685 | } | |
686 | name.setTo(shortID, -1); | |
687 | } | |
688 | noOffsetFormatFallback = TRUE; | |
689 | break; | |
690 | ||
691 | case UTZFMT_STYLE_EXEMPLAR_LOCATION: | |
692 | formatExemplarLocation(tz, name); | |
693 | noOffsetFormatFallback = TRUE; | |
694 | break; | |
695 | ||
51004dcb | 696 | default: |
4388f060 A |
697 | // will be handled below |
698 | break; | |
699 | } | |
700 | ||
57a6839d | 701 | if (name.isEmpty() && !noOffsetFormatFallback) { |
4388f060 A |
702 | UErrorCode status = U_ZERO_ERROR; |
703 | int32_t rawOffset, dstOffset; | |
704 | tz.getOffset(date, FALSE, rawOffset, dstOffset, status); | |
51004dcb | 705 | int32_t offset = rawOffset + dstOffset; |
4388f060 A |
706 | if (U_SUCCESS(status)) { |
707 | switch (style) { | |
51004dcb A |
708 | case UTZFMT_STYLE_GENERIC_LOCATION: |
709 | case UTZFMT_STYLE_GENERIC_LONG: | |
710 | case UTZFMT_STYLE_SPECIFIC_LONG: | |
711 | case UTZFMT_STYLE_LOCALIZED_GMT: | |
712 | formatOffsetLocalizedGMT(offset, name, status); | |
713 | break; | |
714 | ||
715 | case UTZFMT_STYLE_GENERIC_SHORT: | |
716 | case UTZFMT_STYLE_SPECIFIC_SHORT: | |
717 | case UTZFMT_STYLE_LOCALIZED_GMT_SHORT: | |
718 | formatOffsetShortLocalizedGMT(offset, name, status); | |
4388f060 | 719 | break; |
51004dcb A |
720 | |
721 | case UTZFMT_STYLE_ISO_BASIC_SHORT: | |
722 | formatOffsetISO8601Basic(offset, TRUE, TRUE, TRUE, name, status); | |
4388f060 | 723 | break; |
51004dcb A |
724 | |
725 | case UTZFMT_STYLE_ISO_BASIC_LOCAL_SHORT: | |
726 | formatOffsetISO8601Basic(offset, FALSE, TRUE, TRUE, name, status); | |
727 | break; | |
728 | ||
729 | case UTZFMT_STYLE_ISO_BASIC_FIXED: | |
730 | formatOffsetISO8601Basic(offset, TRUE, FALSE, TRUE, name, status); | |
731 | break; | |
732 | ||
733 | case UTZFMT_STYLE_ISO_BASIC_LOCAL_FIXED: | |
734 | formatOffsetISO8601Basic(offset, FALSE, FALSE, TRUE, name, status); | |
735 | break; | |
736 | ||
737 | case UTZFMT_STYLE_ISO_EXTENDED_FIXED: | |
738 | formatOffsetISO8601Extended(offset, TRUE, FALSE, TRUE, name, status); | |
739 | break; | |
740 | ||
741 | case UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FIXED: | |
742 | formatOffsetISO8601Extended(offset, FALSE, FALSE, TRUE, name, status); | |
743 | break; | |
744 | ||
745 | case UTZFMT_STYLE_ISO_BASIC_FULL: | |
746 | formatOffsetISO8601Basic(offset, TRUE, FALSE, FALSE, name, status); | |
747 | break; | |
748 | ||
749 | case UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL: | |
750 | formatOffsetISO8601Basic(offset, FALSE, FALSE, FALSE, name, status); | |
751 | break; | |
752 | ||
753 | case UTZFMT_STYLE_ISO_EXTENDED_FULL: | |
754 | formatOffsetISO8601Extended(offset, TRUE, FALSE, FALSE, name, status); | |
755 | break; | |
756 | ||
757 | case UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FULL: | |
758 | formatOffsetISO8601Extended(offset, FALSE, FALSE, FALSE, name, status); | |
759 | break; | |
760 | ||
57a6839d A |
761 | default: |
762 | // UTZFMT_STYLE_ZONE_ID, UTZFMT_STYLE_ZONE_ID_SHORT, UTZFMT_STYLE_EXEMPLAR_LOCATION | |
763 | break; | |
4388f060 | 764 | } |
57a6839d | 765 | |
4388f060 A |
766 | if (timeType) { |
767 | *timeType = (dstOffset != 0) ? UTZFMT_TIME_TYPE_DAYLIGHT : UTZFMT_TIME_TYPE_STANDARD; | |
768 | } | |
769 | } | |
4388f060 A |
770 | } |
771 | ||
772 | return name; | |
773 | } | |
774 | ||
775 | UnicodeString& | |
776 | TimeZoneFormat::format(const Formattable& obj, UnicodeString& appendTo, | |
777 | FieldPosition& pos, UErrorCode& status) const { | |
778 | if (U_FAILURE(status)) { | |
779 | return appendTo; | |
780 | } | |
781 | UDate date = Calendar::getNow(); | |
782 | if (obj.getType() == Formattable::kObject) { | |
783 | const UObject* formatObj = obj.getObject(); | |
784 | const TimeZone* tz = dynamic_cast<const TimeZone*>(formatObj); | |
785 | if (tz == NULL) { | |
786 | const Calendar* cal = dynamic_cast<const Calendar*>(formatObj); | |
787 | if (cal != NULL) { | |
788 | tz = &cal->getTimeZone(); | |
789 | date = cal->getTime(status); | |
790 | } | |
791 | } | |
792 | if (tz != NULL) { | |
793 | int32_t rawOffset, dstOffset; | |
794 | tz->getOffset(date, FALSE, rawOffset, dstOffset, status); | |
0f5d89e8 | 795 | UChar buf[ZONE_NAME_U16_MAX]; |
b331163b | 796 | UnicodeString result(buf, 0, UPRV_LENGTHOF(buf)); |
4388f060 A |
797 | formatOffsetLocalizedGMT(rawOffset + dstOffset, result, status); |
798 | if (U_SUCCESS(status)) { | |
799 | appendTo.append(result); | |
800 | if (pos.getField() == UDAT_TIMEZONE_FIELD) { | |
801 | pos.setBeginIndex(0); | |
802 | pos.setEndIndex(result.length()); | |
803 | } | |
804 | } | |
805 | } | |
806 | } | |
807 | return appendTo; | |
808 | } | |
809 | ||
810 | TimeZone* | |
811 | TimeZoneFormat::parse(UTimeZoneFormatStyle style, const UnicodeString& text, ParsePosition& pos, | |
812 | UTimeZoneFormatTimeType* timeType /*= NULL*/) const { | |
813 | return parse(style, text, pos, getDefaultParseOptions(), timeType); | |
814 | } | |
815 | ||
816 | TimeZone* | |
817 | TimeZoneFormat::parse(UTimeZoneFormatStyle style, const UnicodeString& text, ParsePosition& pos, | |
818 | int32_t parseOptions, UTimeZoneFormatTimeType* timeType /* = NULL */) const { | |
819 | if (timeType) { | |
820 | *timeType = UTZFMT_TIME_TYPE_UNKNOWN; | |
821 | } | |
822 | ||
823 | int32_t startIdx = pos.getIndex(); | |
824 | int32_t maxPos = text.length(); | |
825 | int32_t offset; | |
826 | ||
51004dcb A |
827 | // Styles using localized GMT format as fallback |
828 | UBool fallbackLocalizedGMT = | |
829 | (style == UTZFMT_STYLE_SPECIFIC_LONG || style == UTZFMT_STYLE_GENERIC_LONG || style == UTZFMT_STYLE_GENERIC_LOCATION); | |
830 | UBool fallbackShortLocalizedGMT = | |
831 | (style == UTZFMT_STYLE_SPECIFIC_SHORT || style == UTZFMT_STYLE_GENERIC_SHORT); | |
4388f060 | 832 | |
51004dcb | 833 | int32_t evaluated = 0; // bit flags representing already evaluated styles |
4388f060 A |
834 | ParsePosition tmpPos(startIdx); |
835 | ||
836 | int32_t parsedOffset = UNKNOWN_OFFSET; // stores successfully parsed offset for later use | |
837 | int32_t parsedPos = -1; // stores successfully parsed offset position for later use | |
838 | ||
839 | // Try localized GMT format first if necessary | |
51004dcb | 840 | if (fallbackLocalizedGMT || fallbackShortLocalizedGMT) { |
4388f060 | 841 | UBool hasDigitOffset = FALSE; |
51004dcb | 842 | offset = parseOffsetLocalizedGMT(text, tmpPos, fallbackShortLocalizedGMT, &hasDigitOffset); |
4388f060 A |
843 | if (tmpPos.getErrorIndex() == -1) { |
844 | // Even when the input text was successfully parsed as a localized GMT format text, | |
845 | // we may still need to evaluate the specified style if - | |
846 | // 1) GMT zero format was used, and | |
847 | // 2) The input text was not completely processed | |
848 | if (tmpPos.getIndex() == maxPos || hasDigitOffset) { | |
849 | pos.setIndex(tmpPos.getIndex()); | |
850 | return createTimeZoneForOffset(offset); | |
851 | } | |
852 | parsedOffset = offset; | |
853 | parsedPos = tmpPos.getIndex(); | |
854 | } | |
51004dcb A |
855 | // Note: For now, no distinction between long/short localized GMT format in the parser. |
856 | // This might be changed in future. | |
857 | // evaluated |= (fallbackLocalizedGMT ? STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT] : STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT_SHORT]); | |
858 | evaluated |= STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT] | STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT_SHORT]; | |
4388f060 A |
859 | } |
860 | ||
861 | UErrorCode status = U_ZERO_ERROR; | |
b331163b A |
862 | UChar tzIDBuf[32]; |
863 | UnicodeString tzID(tzIDBuf, 0, UPRV_LENGTHOF(tzIDBuf)); | |
864 | ||
865 | UBool parseTZDBAbbrev = ((parseOptions & UTZFMT_PARSE_OPTION_TZ_DATABASE_ABBREVIATIONS) != 0); | |
4388f060 A |
866 | |
867 | // Try the specified style | |
868 | switch (style) { | |
51004dcb | 869 | case UTZFMT_STYLE_LOCALIZED_GMT: |
4388f060 | 870 | { |
51004dcb A |
871 | tmpPos.setIndex(startIdx); |
872 | tmpPos.setErrorIndex(-1); | |
873 | ||
874 | offset = parseOffsetLocalizedGMT(text, tmpPos); | |
4388f060 A |
875 | if (tmpPos.getErrorIndex() == -1) { |
876 | pos.setIndex(tmpPos.getIndex()); | |
877 | return createTimeZoneForOffset(offset); | |
878 | } | |
4388f060 | 879 | |
51004dcb A |
880 | // Note: For now, no distinction between long/short localized GMT format in the parser. |
881 | // This might be changed in future. | |
882 | evaluated |= STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT_SHORT]; | |
883 | ||
884 | break; | |
885 | } | |
886 | case UTZFMT_STYLE_LOCALIZED_GMT_SHORT: | |
4388f060 | 887 | { |
51004dcb A |
888 | tmpPos.setIndex(startIdx); |
889 | tmpPos.setErrorIndex(-1); | |
890 | ||
891 | offset = parseOffsetShortLocalizedGMT(text, tmpPos); | |
4388f060 A |
892 | if (tmpPos.getErrorIndex() == -1) { |
893 | pos.setIndex(tmpPos.getIndex()); | |
894 | return createTimeZoneForOffset(offset); | |
895 | } | |
4388f060 | 896 | |
51004dcb A |
897 | // Note: For now, no distinction between long/short localized GMT format in the parser. |
898 | // This might be changed in future. | |
899 | evaluated |= STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT]; | |
900 | ||
901 | break; | |
902 | } | |
903 | case UTZFMT_STYLE_ISO_BASIC_SHORT: | |
904 | case UTZFMT_STYLE_ISO_BASIC_FIXED: | |
905 | case UTZFMT_STYLE_ISO_BASIC_FULL: | |
906 | case UTZFMT_STYLE_ISO_EXTENDED_FIXED: | |
907 | case UTZFMT_STYLE_ISO_EXTENDED_FULL: | |
4388f060 | 908 | { |
51004dcb A |
909 | tmpPos.setIndex(startIdx); |
910 | tmpPos.setErrorIndex(-1); | |
911 | ||
4388f060 A |
912 | offset = parseOffsetISO8601(text, tmpPos); |
913 | if (tmpPos.getErrorIndex() == -1) { | |
914 | pos.setIndex(tmpPos.getIndex()); | |
915 | return createTimeZoneForOffset(offset); | |
916 | } | |
51004dcb A |
917 | |
918 | break; | |
919 | } | |
920 | ||
921 | case UTZFMT_STYLE_ISO_BASIC_LOCAL_SHORT: | |
922 | case UTZFMT_STYLE_ISO_BASIC_LOCAL_FIXED: | |
923 | case UTZFMT_STYLE_ISO_BASIC_LOCAL_FULL: | |
924 | case UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FIXED: | |
925 | case UTZFMT_STYLE_ISO_EXTENDED_LOCAL_FULL: | |
926 | { | |
927 | tmpPos.setIndex(startIdx); | |
928 | tmpPos.setErrorIndex(-1); | |
929 | ||
930 | // Exclude the case of UTC Indicator "Z" here | |
931 | UBool hasDigitOffset = FALSE; | |
932 | offset = parseOffsetISO8601(text, tmpPos, FALSE, &hasDigitOffset); | |
933 | if (tmpPos.getErrorIndex() == -1 && hasDigitOffset) { | |
934 | pos.setIndex(tmpPos.getIndex()); | |
935 | return createTimeZoneForOffset(offset); | |
936 | } | |
937 | ||
938 | break; | |
4388f060 | 939 | } |
4388f060 A |
940 | |
941 | case UTZFMT_STYLE_SPECIFIC_LONG: | |
942 | case UTZFMT_STYLE_SPECIFIC_SHORT: | |
943 | { | |
944 | // Specific styles | |
945 | int32_t nameTypes = 0; | |
946 | if (style == UTZFMT_STYLE_SPECIFIC_LONG) { | |
947 | nameTypes = (UTZNM_LONG_STANDARD | UTZNM_LONG_DAYLIGHT); | |
948 | } else { | |
949 | U_ASSERT(style == UTZFMT_STYLE_SPECIFIC_SHORT); | |
950 | nameTypes = (UTZNM_SHORT_STANDARD | UTZNM_SHORT_DAYLIGHT); | |
951 | } | |
952 | LocalPointer<TimeZoneNames::MatchInfoCollection> specificMatches(fTimeZoneNames->find(text, startIdx, nameTypes, status)); | |
953 | if (U_FAILURE(status)) { | |
954 | pos.setErrorIndex(startIdx); | |
955 | return NULL; | |
956 | } | |
957 | if (!specificMatches.isNull()) { | |
958 | int32_t matchIdx = -1; | |
959 | int32_t matchPos = -1; | |
960 | for (int32_t i = 0; i < specificMatches->size(); i++) { | |
961 | matchPos = startIdx + specificMatches->getMatchLengthAt(i); | |
962 | if (matchPos > parsedPos) { | |
963 | matchIdx = i; | |
964 | parsedPos = matchPos; | |
965 | } | |
966 | } | |
967 | if (matchIdx >= 0) { | |
968 | if (timeType) { | |
969 | *timeType = getTimeType(specificMatches->getNameTypeAt(matchIdx)); | |
970 | } | |
971 | pos.setIndex(matchPos); | |
972 | getTimeZoneID(specificMatches.getAlias(), matchIdx, tzID); | |
973 | U_ASSERT(!tzID.isEmpty()); | |
974 | return TimeZone::createTimeZone(tzID); | |
975 | } | |
976 | } | |
b331163b A |
977 | |
978 | if (parseTZDBAbbrev && style == UTZFMT_STYLE_SPECIFIC_SHORT) { | |
979 | U_ASSERT((nameTypes & UTZNM_SHORT_STANDARD) != 0); | |
980 | U_ASSERT((nameTypes & UTZNM_SHORT_DAYLIGHT) != 0); | |
981 | ||
982 | const TZDBTimeZoneNames *tzdbTimeZoneNames = getTZDBTimeZoneNames(status); | |
983 | if (U_SUCCESS(status)) { | |
984 | LocalPointer<TimeZoneNames::MatchInfoCollection> tzdbNameMatches( | |
985 | tzdbTimeZoneNames->find(text, startIdx, nameTypes, status)); | |
986 | if (U_FAILURE(status)) { | |
987 | pos.setErrorIndex(startIdx); | |
988 | return NULL; | |
989 | } | |
990 | if (!tzdbNameMatches.isNull()) { | |
991 | int32_t matchIdx = -1; | |
992 | int32_t matchPos = -1; | |
993 | for (int32_t i = 0; i < tzdbNameMatches->size(); i++) { | |
994 | matchPos = startIdx + tzdbNameMatches->getMatchLengthAt(i); | |
995 | if (matchPos > parsedPos) { | |
996 | matchIdx = i; | |
997 | parsedPos = matchPos; | |
998 | } | |
999 | } | |
1000 | if (matchIdx >= 0) { | |
1001 | if (timeType) { | |
1002 | *timeType = getTimeType(tzdbNameMatches->getNameTypeAt(matchIdx)); | |
1003 | } | |
1004 | pos.setIndex(matchPos); | |
1005 | getTimeZoneID(tzdbNameMatches.getAlias(), matchIdx, tzID); | |
1006 | U_ASSERT(!tzID.isEmpty()); | |
1007 | return TimeZone::createTimeZone(tzID); | |
1008 | } | |
1009 | } | |
1010 | } | |
1011 | } | |
51004dcb | 1012 | break; |
4388f060 | 1013 | } |
4388f060 A |
1014 | case UTZFMT_STYLE_GENERIC_LONG: |
1015 | case UTZFMT_STYLE_GENERIC_SHORT: | |
1016 | case UTZFMT_STYLE_GENERIC_LOCATION: | |
1017 | { | |
1018 | int32_t genericNameTypes = 0; | |
1019 | switch (style) { | |
1020 | case UTZFMT_STYLE_GENERIC_LOCATION: | |
1021 | genericNameTypes = UTZGNM_LOCATION; | |
1022 | break; | |
1023 | ||
1024 | case UTZFMT_STYLE_GENERIC_LONG: | |
1025 | genericNameTypes = UTZGNM_LONG | UTZGNM_LOCATION; | |
1026 | break; | |
1027 | ||
1028 | case UTZFMT_STYLE_GENERIC_SHORT: | |
1029 | genericNameTypes = UTZGNM_SHORT | UTZGNM_LOCATION; | |
1030 | break; | |
1031 | ||
1032 | default: | |
3d1f044b | 1033 | UPRV_UNREACHABLE; |
4388f060 A |
1034 | } |
1035 | ||
1036 | int32_t len = 0; | |
51004dcb | 1037 | UTimeZoneFormatTimeType tt = UTZFMT_TIME_TYPE_UNKNOWN; |
4388f060 A |
1038 | const TimeZoneGenericNames *gnames = getTimeZoneGenericNames(status); |
1039 | if (U_SUCCESS(status)) { | |
51004dcb | 1040 | len = gnames->findBestMatch(text, startIdx, genericNameTypes, tzID, tt, status); |
4388f060 A |
1041 | } |
1042 | if (U_FAILURE(status)) { | |
1043 | pos.setErrorIndex(startIdx); | |
1044 | return NULL; | |
1045 | } | |
1046 | if (len > 0) { | |
1047 | // Found a match | |
1048 | if (timeType) { | |
51004dcb | 1049 | *timeType = tt; |
4388f060 A |
1050 | } |
1051 | pos.setIndex(startIdx + len); | |
1052 | U_ASSERT(!tzID.isEmpty()); | |
1053 | return TimeZone::createTimeZone(tzID); | |
1054 | } | |
51004dcb A |
1055 | |
1056 | break; | |
1057 | } | |
1058 | case UTZFMT_STYLE_ZONE_ID: | |
1059 | { | |
1060 | tmpPos.setIndex(startIdx); | |
1061 | tmpPos.setErrorIndex(-1); | |
1062 | ||
1063 | parseZoneID(text, tmpPos, tzID); | |
1064 | if (tmpPos.getErrorIndex() == -1) { | |
1065 | pos.setIndex(tmpPos.getIndex()); | |
1066 | return TimeZone::createTimeZone(tzID); | |
1067 | } | |
1068 | break; | |
1069 | } | |
1070 | case UTZFMT_STYLE_ZONE_ID_SHORT: | |
1071 | { | |
1072 | tmpPos.setIndex(startIdx); | |
1073 | tmpPos.setErrorIndex(-1); | |
1074 | ||
1075 | parseShortZoneID(text, tmpPos, tzID); | |
1076 | if (tmpPos.getErrorIndex() == -1) { | |
1077 | pos.setIndex(tmpPos.getIndex()); | |
1078 | return TimeZone::createTimeZone(tzID); | |
1079 | } | |
1080 | break; | |
1081 | } | |
1082 | case UTZFMT_STYLE_EXEMPLAR_LOCATION: | |
1083 | { | |
1084 | tmpPos.setIndex(startIdx); | |
1085 | tmpPos.setErrorIndex(-1); | |
1086 | ||
1087 | parseExemplarLocation(text, tmpPos, tzID); | |
1088 | if (tmpPos.getErrorIndex() == -1) { | |
1089 | pos.setIndex(tmpPos.getIndex()); | |
1090 | return TimeZone::createTimeZone(tzID); | |
1091 | } | |
1092 | break; | |
4388f060 | 1093 | } |
4388f060 | 1094 | } |
51004dcb | 1095 | evaluated |= STYLE_PARSE_FLAGS[style]; |
4388f060 A |
1096 | |
1097 | ||
1098 | if (parsedPos > startIdx) { | |
1099 | // When the specified style is one of SPECIFIC_XXX or GENERIC_XXX, we tried to parse the input | |
1100 | // as localized GMT format earlier. If parsedOffset is positive, it means it was successfully | |
1101 | // parsed as localized GMT format, but offset digits were not detected (more specifically, GMT | |
1102 | // zero format). Then, it tried to find a match within the set of display names, but could not | |
1103 | // find a match. At this point, we can safely assume the input text contains the localized | |
1104 | // GMT format. | |
1105 | U_ASSERT(parsedOffset != UNKNOWN_OFFSET); | |
1106 | pos.setIndex(parsedPos); | |
1107 | return createTimeZoneForOffset(parsedOffset); | |
1108 | } | |
1109 | ||
1110 | // Failed to parse the input text as the time zone format in the specified style. | |
1111 | // Check the longest match among other styles below. | |
b331163b A |
1112 | UChar parsedIDBuf[32]; |
1113 | UnicodeString parsedID(parsedIDBuf, 0, UPRV_LENGTHOF(parsedIDBuf)); | |
51004dcb A |
1114 | UTimeZoneFormatTimeType parsedTimeType = UTZFMT_TIME_TYPE_UNKNOWN; |
1115 | ||
4388f060 A |
1116 | U_ASSERT(parsedPos < 0); |
1117 | U_ASSERT(parsedOffset == UNKNOWN_OFFSET); | |
4388f060 A |
1118 | |
1119 | // ISO 8601 | |
51004dcb A |
1120 | if (parsedPos < maxPos && |
1121 | ((evaluated & ISO_Z_STYLE_FLAG) == 0 || (evaluated & ISO_LOCAL_STYLE_FLAG) == 0)) { | |
1122 | tmpPos.setIndex(startIdx); | |
1123 | tmpPos.setErrorIndex(-1); | |
1124 | ||
4388f060 A |
1125 | UBool hasDigitOffset = FALSE; |
1126 | offset = parseOffsetISO8601(text, tmpPos, FALSE, &hasDigitOffset); | |
1127 | if (tmpPos.getErrorIndex() == -1) { | |
1128 | if (tmpPos.getIndex() == maxPos || hasDigitOffset) { | |
1129 | pos.setIndex(tmpPos.getIndex()); | |
1130 | return createTimeZoneForOffset(offset); | |
1131 | } | |
1132 | // Note: When ISO 8601 format contains offset digits, it should not | |
51004dcb A |
1133 | // collide with other formats. However, ISO 8601 UTC format "Z" (single letter) |
1134 | // may collide with other names. In this case, we need to evaluate other names. | |
1135 | if (parsedPos < tmpPos.getIndex()) { | |
1136 | parsedOffset = offset; | |
1137 | parsedID.setToBogus(); | |
1138 | parsedTimeType = UTZFMT_TIME_TYPE_UNKNOWN; | |
1139 | parsedPos = tmpPos.getIndex(); | |
1140 | U_ASSERT(parsedPos == startIdx + 1); // only when "Z" is used | |
1141 | } | |
4388f060 | 1142 | } |
51004dcb A |
1143 | } |
1144 | ||
1145 | // Localized GMT format | |
1146 | if (parsedPos < maxPos && | |
1147 | (evaluated & STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT]) == 0) { | |
4388f060 A |
1148 | tmpPos.setIndex(startIdx); |
1149 | tmpPos.setErrorIndex(-1); | |
4388f060 | 1150 | |
51004dcb A |
1151 | UBool hasDigitOffset = FALSE; |
1152 | offset = parseOffsetLocalizedGMT(text, tmpPos, FALSE, &hasDigitOffset); | |
1153 | if (tmpPos.getErrorIndex() == -1) { | |
1154 | if (tmpPos.getIndex() == maxPos || hasDigitOffset) { | |
1155 | pos.setIndex(tmpPos.getIndex()); | |
1156 | return createTimeZoneForOffset(offset); | |
1157 | } | |
1158 | // Evaluate other names - see the comment earlier in this method. | |
1159 | if (parsedPos < tmpPos.getIndex()) { | |
1160 | parsedOffset = offset; | |
1161 | parsedID.setToBogus(); | |
1162 | parsedTimeType = UTZFMT_TIME_TYPE_UNKNOWN; | |
1163 | parsedPos = tmpPos.getIndex(); | |
1164 | } | |
1165 | } | |
1166 | } | |
4388f060 | 1167 | |
51004dcb A |
1168 | if (parsedPos < maxPos && |
1169 | (evaluated & STYLE_PARSE_FLAGS[UTZFMT_STYLE_LOCALIZED_GMT_SHORT]) == 0) { | |
1170 | tmpPos.setIndex(startIdx); | |
1171 | tmpPos.setErrorIndex(-1); | |
4388f060 | 1172 | |
4388f060 | 1173 | UBool hasDigitOffset = FALSE; |
51004dcb | 1174 | offset = parseOffsetLocalizedGMT(text, tmpPos, TRUE, &hasDigitOffset); |
4388f060 A |
1175 | if (tmpPos.getErrorIndex() == -1) { |
1176 | if (tmpPos.getIndex() == maxPos || hasDigitOffset) { | |
1177 | pos.setIndex(tmpPos.getIndex()); | |
1178 | return createTimeZoneForOffset(offset); | |
1179 | } | |
1180 | // Evaluate other names - see the comment earlier in this method. | |
51004dcb A |
1181 | if (parsedPos < tmpPos.getIndex()) { |
1182 | parsedOffset = offset; | |
1183 | parsedID.setToBogus(); | |
1184 | parsedTimeType = UTZFMT_TIME_TYPE_UNKNOWN; | |
1185 | parsedPos = tmpPos.getIndex(); | |
1186 | } | |
4388f060 A |
1187 | } |
1188 | } | |
1189 | ||
51004dcb | 1190 | // When ParseOption.ALL_STYLES is available, we also try to look all possible display names and IDs. |
4388f060 A |
1191 | // For example, when style is GENERIC_LONG, "EST" (SPECIFIC_SHORT) is never |
1192 | // used for America/New_York. With parseAllStyles true, this code parses "EST" | |
1193 | // as America/New_York. | |
1194 | ||
1195 | // Note: Adding all possible names into the trie used by the implementation is quite heavy operation, | |
1196 | // which we want to avoid normally (note that we cache the trie, so this is applicable to the | |
1197 | // first time only as long as the cache does not expire). | |
51004dcb | 1198 | |
4388f060 | 1199 | if (parseOptions & UTZFMT_PARSE_OPTION_ALL_STYLES) { |
51004dcb A |
1200 | // Try all specific names and exemplar location names |
1201 | if (parsedPos < maxPos) { | |
1202 | LocalPointer<TimeZoneNames::MatchInfoCollection> specificMatches(fTimeZoneNames->find(text, startIdx, ALL_SIMPLE_NAME_TYPES, status)); | |
1203 | if (U_FAILURE(status)) { | |
1204 | pos.setErrorIndex(startIdx); | |
1205 | return NULL; | |
1206 | } | |
1207 | int32_t specificMatchIdx = -1; | |
4388f060 | 1208 | int32_t matchPos = -1; |
51004dcb A |
1209 | if (!specificMatches.isNull()) { |
1210 | for (int32_t i = 0; i < specificMatches->size(); i++) { | |
1211 | if (startIdx + specificMatches->getMatchLengthAt(i) > matchPos) { | |
1212 | specificMatchIdx = i; | |
1213 | matchPos = startIdx + specificMatches->getMatchLengthAt(i); | |
1214 | } | |
4388f060 A |
1215 | } |
1216 | } | |
51004dcb A |
1217 | if (parsedPos < matchPos) { |
1218 | U_ASSERT(specificMatchIdx >= 0); | |
1219 | parsedPos = matchPos; | |
1220 | getTimeZoneID(specificMatches.getAlias(), specificMatchIdx, parsedID); | |
1221 | parsedTimeType = getTimeType(specificMatches->getNameTypeAt(specificMatchIdx)); | |
1222 | parsedOffset = UNKNOWN_OFFSET; | |
1223 | } | |
4388f060 | 1224 | } |
b331163b A |
1225 | if (parseTZDBAbbrev && parsedPos < maxPos && (evaluated & STYLE_PARSE_FLAGS[UTZFMT_STYLE_SPECIFIC_SHORT]) == 0) { |
1226 | const TZDBTimeZoneNames *tzdbTimeZoneNames = getTZDBTimeZoneNames(status); | |
1227 | if (U_SUCCESS(status)) { | |
1228 | LocalPointer<TimeZoneNames::MatchInfoCollection> tzdbNameMatches( | |
1229 | tzdbTimeZoneNames->find(text, startIdx, ALL_SIMPLE_NAME_TYPES, status)); | |
1230 | if (U_FAILURE(status)) { | |
1231 | pos.setErrorIndex(startIdx); | |
1232 | return NULL; | |
1233 | } | |
1234 | int32_t tzdbNameMatchIdx = -1; | |
1235 | int32_t matchPos = -1; | |
1236 | if (!tzdbNameMatches.isNull()) { | |
1237 | for (int32_t i = 0; i < tzdbNameMatches->size(); i++) { | |
1238 | if (startIdx + tzdbNameMatches->getMatchLengthAt(i) > matchPos) { | |
1239 | tzdbNameMatchIdx = i; | |
1240 | matchPos = startIdx + tzdbNameMatches->getMatchLengthAt(i); | |
1241 | } | |
1242 | } | |
1243 | } | |
1244 | if (parsedPos < matchPos) { | |
1245 | U_ASSERT(tzdbNameMatchIdx >= 0); | |
1246 | parsedPos = matchPos; | |
1247 | getTimeZoneID(tzdbNameMatches.getAlias(), tzdbNameMatchIdx, parsedID); | |
1248 | parsedTimeType = getTimeType(tzdbNameMatches->getNameTypeAt(tzdbNameMatchIdx)); | |
1249 | parsedOffset = UNKNOWN_OFFSET; | |
1250 | } | |
1251 | } | |
1252 | } | |
51004dcb | 1253 | // Try generic names |
4388f060 | 1254 | if (parsedPos < maxPos) { |
51004dcb A |
1255 | int32_t genMatchLen = -1; |
1256 | UTimeZoneFormatTimeType tt = UTZFMT_TIME_TYPE_UNKNOWN; | |
1257 | ||
4388f060 A |
1258 | const TimeZoneGenericNames *gnames = getTimeZoneGenericNames(status); |
1259 | if (U_SUCCESS(status)) { | |
51004dcb | 1260 | genMatchLen = gnames->findBestMatch(text, startIdx, ALL_GENERIC_NAME_TYPES, tzID, tt, status); |
4388f060 A |
1261 | } |
1262 | if (U_FAILURE(status)) { | |
1263 | pos.setErrorIndex(startIdx); | |
1264 | return NULL; | |
1265 | } | |
51004dcb | 1266 | |
b331163b | 1267 | if (genMatchLen > 0 && parsedPos < startIdx + genMatchLen) { |
51004dcb A |
1268 | parsedPos = startIdx + genMatchLen; |
1269 | parsedID.setTo(tzID); | |
1270 | parsedTimeType = tt; | |
1271 | parsedOffset = UNKNOWN_OFFSET; | |
1272 | } | |
4388f060 | 1273 | } |
51004dcb A |
1274 | |
1275 | // Try time zone ID | |
1276 | if (parsedPos < maxPos && (evaluated & STYLE_PARSE_FLAGS[UTZFMT_STYLE_ZONE_ID]) == 0) { | |
1277 | tmpPos.setIndex(startIdx); | |
1278 | tmpPos.setErrorIndex(-1); | |
1279 | ||
1280 | parseZoneID(text, tmpPos, tzID); | |
1281 | if (tmpPos.getErrorIndex() == -1 && parsedPos < tmpPos.getIndex()) { | |
1282 | parsedPos = tmpPos.getIndex(); | |
1283 | parsedID.setTo(tzID); | |
1284 | parsedTimeType = UTZFMT_TIME_TYPE_UNKNOWN; | |
1285 | parsedOffset = UNKNOWN_OFFSET; | |
4388f060 | 1286 | } |
51004dcb A |
1287 | } |
1288 | // Try short time zone ID | |
1289 | if (parsedPos < maxPos && (evaluated & STYLE_PARSE_FLAGS[UTZFMT_STYLE_ZONE_ID]) == 0) { | |
1290 | tmpPos.setIndex(startIdx); | |
1291 | tmpPos.setErrorIndex(-1); | |
1292 | ||
1293 | parseShortZoneID(text, tmpPos, tzID); | |
1294 | if (tmpPos.getErrorIndex() == -1 && parsedPos < tmpPos.getIndex()) { | |
1295 | parsedPos = tmpPos.getIndex(); | |
1296 | parsedID.setTo(tzID); | |
1297 | parsedTimeType = UTZFMT_TIME_TYPE_UNKNOWN; | |
1298 | parsedOffset = UNKNOWN_OFFSET; | |
4388f060 | 1299 | } |
4388f060 A |
1300 | } |
1301 | } | |
1302 | ||
1303 | if (parsedPos > startIdx) { | |
51004dcb A |
1304 | // Parsed successfully |
1305 | TimeZone* parsedTZ; | |
1306 | if (parsedID.length() > 0) { | |
1307 | parsedTZ = TimeZone::createTimeZone(parsedID); | |
1308 | } else { | |
1309 | U_ASSERT(parsedOffset != UNKNOWN_OFFSET); | |
1310 | parsedTZ = createTimeZoneForOffset(parsedOffset); | |
1311 | } | |
1312 | if (timeType) { | |
1313 | *timeType = parsedTimeType; | |
1314 | } | |
4388f060 | 1315 | pos.setIndex(parsedPos); |
51004dcb | 1316 | return parsedTZ; |
4388f060 A |
1317 | } |
1318 | ||
1319 | pos.setErrorIndex(startIdx); | |
1320 | return NULL; | |
1321 | } | |
1322 | ||
1323 | void | |
1324 | TimeZoneFormat::parseObject(const UnicodeString& source, Formattable& result, | |
1325 | ParsePosition& parse_pos) const { | |
1326 | result.adoptObject(parse(UTZFMT_STYLE_GENERIC_LOCATION, source, parse_pos, UTZFMT_PARSE_OPTION_ALL_STYLES)); | |
1327 | } | |
1328 | ||
1329 | ||
1330 | // ------------------------------------------------------------------ | |
1331 | // Private zone name format/parse implementation | |
1332 | ||
1333 | UnicodeString& | |
1334 | TimeZoneFormat::formatGeneric(const TimeZone& tz, int32_t genType, UDate date, UnicodeString& name) const { | |
1335 | UErrorCode status = U_ZERO_ERROR; | |
1336 | const TimeZoneGenericNames* gnames = getTimeZoneGenericNames(status); | |
1337 | if (U_FAILURE(status)) { | |
1338 | name.setToBogus(); | |
1339 | return name; | |
1340 | } | |
1341 | ||
1342 | if (genType == UTZGNM_LOCATION) { | |
1343 | const UChar* canonicalID = ZoneMeta::getCanonicalCLDRID(tz); | |
1344 | if (canonicalID == NULL) { | |
1345 | name.setToBogus(); | |
1346 | return name; | |
1347 | } | |
b331163b | 1348 | return gnames->getGenericLocationName(UnicodeString(TRUE, canonicalID, -1), name); |
4388f060 A |
1349 | } |
1350 | return gnames->getDisplayName(tz, (UTimeZoneGenericNameType)genType, date, name); | |
1351 | } | |
1352 | ||
1353 | UnicodeString& | |
1354 | TimeZoneFormat::formatSpecific(const TimeZone& tz, UTimeZoneNameType stdType, UTimeZoneNameType dstType, | |
1355 | UDate date, UnicodeString& name, UTimeZoneFormatTimeType *timeType) const { | |
1356 | if (fTimeZoneNames == NULL) { | |
1357 | name.setToBogus(); | |
1358 | return name; | |
1359 | } | |
1360 | ||
1361 | UErrorCode status = U_ZERO_ERROR; | |
1362 | UBool isDaylight = tz.inDaylightTime(date, status); | |
1363 | const UChar* canonicalID = ZoneMeta::getCanonicalCLDRID(tz); | |
1364 | ||
1365 | if (U_FAILURE(status) || canonicalID == NULL) { | |
1366 | name.setToBogus(); | |
1367 | return name; | |
1368 | } | |
1369 | ||
1370 | if (isDaylight) { | |
b331163b | 1371 | fTimeZoneNames->getDisplayName(UnicodeString(TRUE, canonicalID, -1), dstType, date, name); |
4388f060 | 1372 | } else { |
b331163b | 1373 | fTimeZoneNames->getDisplayName(UnicodeString(TRUE, canonicalID, -1), stdType, date, name); |
4388f060 A |
1374 | } |
1375 | ||
1376 | if (timeType && !name.isEmpty()) { | |
1377 | *timeType = isDaylight ? UTZFMT_TIME_TYPE_DAYLIGHT : UTZFMT_TIME_TYPE_STANDARD; | |
1378 | } | |
1379 | return name; | |
1380 | } | |
1381 | ||
1382 | const TimeZoneGenericNames* | |
1383 | TimeZoneFormat::getTimeZoneGenericNames(UErrorCode& status) const { | |
1384 | if (U_FAILURE(status)) { | |
1385 | return NULL; | |
1386 | } | |
1387 | ||
3d1f044b | 1388 | umtx_lock(gLock()); |
57a6839d | 1389 | if (fTimeZoneGenericNames == NULL) { |
4388f060 | 1390 | TimeZoneFormat *nonConstThis = const_cast<TimeZoneFormat *>(this); |
57a6839d | 1391 | nonConstThis->fTimeZoneGenericNames = TimeZoneGenericNames::createInstance(fLocale, status); |
4388f060 | 1392 | } |
3d1f044b | 1393 | umtx_unlock(gLock()); |
4388f060 A |
1394 | |
1395 | return fTimeZoneGenericNames; | |
1396 | } | |
1397 | ||
b331163b A |
1398 | const TZDBTimeZoneNames* |
1399 | TimeZoneFormat::getTZDBTimeZoneNames(UErrorCode& status) const { | |
1400 | if (U_FAILURE(status)) { | |
1401 | return NULL; | |
1402 | } | |
1403 | ||
3d1f044b | 1404 | umtx_lock(gLock()); |
b331163b A |
1405 | if (fTZDBTimeZoneNames == NULL) { |
1406 | TZDBTimeZoneNames *tzdbNames = new TZDBTimeZoneNames(fLocale); | |
1407 | if (tzdbNames == NULL) { | |
1408 | status = U_MEMORY_ALLOCATION_ERROR; | |
1409 | } else { | |
1410 | TimeZoneFormat *nonConstThis = const_cast<TimeZoneFormat *>(this); | |
1411 | nonConstThis->fTZDBTimeZoneNames = tzdbNames; | |
1412 | } | |
1413 | } | |
3d1f044b | 1414 | umtx_unlock(gLock()); |
b331163b A |
1415 | |
1416 | return fTZDBTimeZoneNames; | |
1417 | } | |
1418 | ||
51004dcb A |
1419 | UnicodeString& |
1420 | TimeZoneFormat::formatExemplarLocation(const TimeZone& tz, UnicodeString& name) const { | |
0f5d89e8 | 1421 | UChar locationBuf[ZONE_NAME_U16_MAX]; |
b331163b | 1422 | UnicodeString location(locationBuf, 0, UPRV_LENGTHOF(locationBuf)); |
51004dcb A |
1423 | const UChar* canonicalID = ZoneMeta::getCanonicalCLDRID(tz); |
1424 | ||
1425 | if (canonicalID) { | |
b331163b | 1426 | fTimeZoneNames->getExemplarLocationName(UnicodeString(TRUE, canonicalID, -1), location); |
51004dcb A |
1427 | } |
1428 | if (location.length() > 0) { | |
1429 | name.setTo(location); | |
1430 | } else { | |
1431 | // Use "unknown" location | |
b331163b | 1432 | fTimeZoneNames->getExemplarLocationName(UnicodeString(TRUE, UNKNOWN_ZONE_ID, -1), location); |
51004dcb A |
1433 | if (location.length() > 0) { |
1434 | name.setTo(location); | |
1435 | } else { | |
1436 | // last resort | |
1437 | name.setTo(UNKNOWN_LOCATION, -1); | |
1438 | } | |
1439 | } | |
1440 | return name; | |
1441 | } | |
1442 | ||
1443 | ||
4388f060 A |
1444 | // ------------------------------------------------------------------ |
1445 | // Zone offset format and parse | |
1446 | ||
1447 | UnicodeString& | |
51004dcb A |
1448 | TimeZoneFormat::formatOffsetISO8601Basic(int32_t offset, UBool useUtcIndicator, UBool isShort, UBool ignoreSeconds, |
1449 | UnicodeString& result, UErrorCode& status) const { | |
1450 | return formatOffsetISO8601(offset, TRUE, useUtcIndicator, isShort, ignoreSeconds, result, status); | |
1451 | } | |
1452 | ||
1453 | UnicodeString& | |
1454 | TimeZoneFormat::formatOffsetISO8601Extended(int32_t offset, UBool useUtcIndicator, UBool isShort, UBool ignoreSeconds, | |
1455 | UnicodeString& result, UErrorCode& status) const { | |
1456 | return formatOffsetISO8601(offset, FALSE, useUtcIndicator, isShort, ignoreSeconds, result, status); | |
1457 | } | |
1458 | ||
1459 | UnicodeString& | |
1460 | TimeZoneFormat::formatOffsetLocalizedGMT(int32_t offset, UnicodeString& result, UErrorCode& status) const { | |
1461 | return formatOffsetLocalizedGMT(offset, FALSE, result, status); | |
1462 | } | |
1463 | ||
1464 | UnicodeString& | |
1465 | TimeZoneFormat::formatOffsetShortLocalizedGMT(int32_t offset, UnicodeString& result, UErrorCode& status) const { | |
1466 | return formatOffsetLocalizedGMT(offset, TRUE, result, status); | |
1467 | } | |
1468 | ||
1469 | int32_t | |
1470 | TimeZoneFormat::parseOffsetISO8601(const UnicodeString& text, ParsePosition& pos) const { | |
1471 | return parseOffsetISO8601(text, pos, FALSE); | |
1472 | } | |
1473 | ||
1474 | int32_t | |
1475 | TimeZoneFormat::parseOffsetLocalizedGMT(const UnicodeString& text, ParsePosition& pos) const { | |
1476 | return parseOffsetLocalizedGMT(text, pos, FALSE, NULL); | |
1477 | } | |
1478 | ||
1479 | int32_t | |
1480 | TimeZoneFormat::parseOffsetShortLocalizedGMT(const UnicodeString& text, ParsePosition& pos) const { | |
1481 | return parseOffsetLocalizedGMT(text, pos, TRUE, NULL); | |
1482 | } | |
1483 | ||
1484 | // ------------------------------------------------------------------ | |
1485 | // Private zone offset format/parse implementation | |
1486 | ||
1487 | UnicodeString& | |
1488 | TimeZoneFormat::formatOffsetISO8601(int32_t offset, UBool isBasic, UBool useUtcIndicator, | |
1489 | UBool isShort, UBool ignoreSeconds, UnicodeString& result, UErrorCode& status) const { | |
4388f060 A |
1490 | if (U_FAILURE(status)) { |
1491 | result.setToBogus(); | |
1492 | return result; | |
1493 | } | |
51004dcb A |
1494 | int32_t absOffset = offset < 0 ? -offset : offset; |
1495 | if (useUtcIndicator && (absOffset < MILLIS_PER_SECOND || (ignoreSeconds && absOffset < MILLIS_PER_MINUTE))) { | |
1496 | result.setTo(ISO8601_UTC); | |
1497 | return result; | |
1498 | } | |
1499 | ||
1500 | OffsetFields minFields = isShort ? FIELDS_H : FIELDS_HM; | |
1501 | OffsetFields maxFields = ignoreSeconds ? FIELDS_HM : FIELDS_HMS; | |
1502 | UChar sep = isBasic ? 0 : ISO8601_SEP; | |
1503 | ||
1504 | // Note: FIELDS_HMS as maxFields is a CLDR/ICU extension. ISO 8601 specification does | |
1505 | // not support seconds field. | |
1506 | ||
1507 | if (absOffset >= MAX_OFFSET) { | |
4388f060 A |
1508 | result.setToBogus(); |
1509 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
1510 | return result; | |
1511 | } | |
1512 | ||
51004dcb A |
1513 | int fields[3]; |
1514 | fields[0] = absOffset / MILLIS_PER_HOUR; | |
1515 | absOffset = absOffset % MILLIS_PER_HOUR; | |
1516 | fields[1] = absOffset / MILLIS_PER_MINUTE; | |
1517 | absOffset = absOffset % MILLIS_PER_MINUTE; | |
1518 | fields[2] = absOffset / MILLIS_PER_SECOND; | |
1519 | ||
1520 | U_ASSERT(fields[0] >= 0 && fields[0] <= MAX_OFFSET_HOUR); | |
1521 | U_ASSERT(fields[1] >= 0 && fields[1] <= MAX_OFFSET_MINUTE); | |
1522 | U_ASSERT(fields[2] >= 0 && fields[2] <= MAX_OFFSET_SECOND); | |
1523 | ||
1524 | int32_t lastIdx = maxFields; | |
1525 | while (lastIdx > minFields) { | |
1526 | if (fields[lastIdx] != 0) { | |
1527 | break; | |
1528 | } | |
1529 | lastIdx--; | |
1530 | } | |
1531 | ||
1532 | UChar sign = PLUS; | |
1533 | if (offset < 0) { | |
1534 | // if all output fields are 0s, do not use negative sign | |
1535 | for (int32_t idx = 0; idx <= lastIdx; idx++) { | |
1536 | if (fields[idx] != 0) { | |
1537 | sign = MINUS; | |
1538 | break; | |
1539 | } | |
1540 | } | |
1541 | } | |
1542 | result.setTo(sign); | |
1543 | ||
1544 | for (int32_t idx = 0; idx <= lastIdx; idx++) { | |
1545 | if (sep && idx != 0) { | |
1546 | result.append(sep); | |
1547 | } | |
1548 | result.append((UChar)(0x0030 + fields[idx]/10)); | |
1549 | result.append((UChar)(0x0030 + fields[idx]%10)); | |
1550 | } | |
1551 | ||
1552 | return result; | |
4388f060 A |
1553 | } |
1554 | ||
1555 | UnicodeString& | |
51004dcb | 1556 | TimeZoneFormat::formatOffsetLocalizedGMT(int32_t offset, UBool isShort, UnicodeString& result, UErrorCode& status) const { |
4388f060 A |
1557 | if (U_FAILURE(status)) { |
1558 | result.setToBogus(); | |
1559 | return result; | |
1560 | } | |
1561 | if (offset <= -MAX_OFFSET || offset >= MAX_OFFSET) { | |
1562 | result.setToBogus(); | |
1563 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
1564 | return result; | |
1565 | } | |
1566 | ||
1567 | if (offset == 0) { | |
1568 | result.setTo(fGMTZeroFormat); | |
1569 | return result; | |
1570 | } | |
1571 | ||
1572 | UBool positive = TRUE; | |
1573 | if (offset < 0) { | |
1574 | offset = -offset; | |
1575 | positive = FALSE; | |
1576 | } | |
1577 | ||
1578 | int32_t offsetH = offset / MILLIS_PER_HOUR; | |
1579 | offset = offset % MILLIS_PER_HOUR; | |
1580 | int32_t offsetM = offset / MILLIS_PER_MINUTE; | |
1581 | offset = offset % MILLIS_PER_MINUTE; | |
1582 | int32_t offsetS = offset / MILLIS_PER_SECOND; | |
1583 | ||
1584 | U_ASSERT(offsetH <= MAX_OFFSET_HOUR && offsetM <= MAX_OFFSET_MINUTE && offsetS <= MAX_OFFSET_SECOND); | |
1585 | ||
1586 | const UVector* offsetPatternItems = NULL; | |
1587 | if (positive) { | |
51004dcb A |
1588 | if (offsetS != 0) { |
1589 | offsetPatternItems = fGMTOffsetPatternItems[UTZFMT_PAT_POSITIVE_HMS]; | |
1590 | } else if (offsetM != 0 || !isShort) { | |
1591 | offsetPatternItems = fGMTOffsetPatternItems[UTZFMT_PAT_POSITIVE_HM]; | |
1592 | } else { | |
1593 | offsetPatternItems = fGMTOffsetPatternItems[UTZFMT_PAT_POSITIVE_H]; | |
1594 | } | |
4388f060 | 1595 | } else { |
51004dcb A |
1596 | if (offsetS != 0) { |
1597 | offsetPatternItems = fGMTOffsetPatternItems[UTZFMT_PAT_NEGATIVE_HMS]; | |
1598 | } else if (offsetM != 0 || !isShort) { | |
1599 | offsetPatternItems = fGMTOffsetPatternItems[UTZFMT_PAT_NEGATIVE_HM]; | |
1600 | } else { | |
1601 | offsetPatternItems = fGMTOffsetPatternItems[UTZFMT_PAT_NEGATIVE_H]; | |
1602 | } | |
4388f060 A |
1603 | } |
1604 | ||
1605 | U_ASSERT(offsetPatternItems != NULL); | |
1606 | ||
1607 | // Building the GMT format string | |
1608 | result.setTo(fGMTPatternPrefix); | |
1609 | ||
1610 | for (int32_t i = 0; i < offsetPatternItems->size(); i++) { | |
1611 | const GMTOffsetField* item = (GMTOffsetField*)offsetPatternItems->elementAt(i); | |
1612 | GMTOffsetField::FieldType type = item->getType(); | |
1613 | ||
1614 | switch (type) { | |
1615 | case GMTOffsetField::TEXT: | |
1616 | result.append(item->getPatternText(), -1); | |
1617 | break; | |
1618 | ||
1619 | case GMTOffsetField::HOUR: | |
51004dcb | 1620 | appendOffsetDigits(result, offsetH, (isShort ? 1 : 2)); |
4388f060 A |
1621 | break; |
1622 | ||
1623 | case GMTOffsetField::MINUTE: | |
51004dcb | 1624 | appendOffsetDigits(result, offsetM, 2); |
4388f060 A |
1625 | break; |
1626 | ||
1627 | case GMTOffsetField::SECOND: | |
51004dcb | 1628 | appendOffsetDigits(result, offsetS, 2); |
4388f060 A |
1629 | break; |
1630 | } | |
1631 | } | |
1632 | ||
1633 | result.append(fGMTPatternSuffix); | |
1634 | return result; | |
1635 | } | |
1636 | ||
4388f060 A |
1637 | int32_t |
1638 | TimeZoneFormat::parseOffsetISO8601(const UnicodeString& text, ParsePosition& pos, UBool extendedOnly, UBool* hasDigitOffset /* = NULL */) const { | |
1639 | if (hasDigitOffset) { | |
1640 | *hasDigitOffset = FALSE; | |
1641 | } | |
1642 | int32_t start = pos.getIndex(); | |
1643 | if (start >= text.length()) { | |
1644 | pos.setErrorIndex(start); | |
1645 | return 0; | |
1646 | } | |
1647 | ||
1648 | UChar firstChar = text.charAt(start); | |
1649 | if (firstChar == ISO8601_UTC || firstChar == (UChar)(ISO8601_UTC + 0x20)) { | |
1650 | // "Z" (or "z") - indicates UTC | |
1651 | pos.setIndex(start + 1); | |
1652 | return 0; | |
1653 | } | |
1654 | ||
1655 | int32_t sign = 1; | |
1656 | if (firstChar == PLUS) { | |
1657 | sign = 1; | |
1658 | } else if (firstChar == MINUS) { | |
1659 | sign = -1; | |
1660 | } else { | |
1661 | // Not an ISO 8601 offset string | |
1662 | pos.setErrorIndex(start); | |
1663 | return 0; | |
1664 | } | |
1665 | ParsePosition posOffset(start + 1); | |
51004dcb | 1666 | int32_t offset = parseAsciiOffsetFields(text, posOffset, ISO8601_SEP, FIELDS_H, FIELDS_HMS); |
4388f060 A |
1667 | if (posOffset.getErrorIndex() == -1 && !extendedOnly && (posOffset.getIndex() - start <= 3)) { |
1668 | // If the text is successfully parsed as extended format with the options above, it can be also parsed | |
1669 | // as basic format. For example, "0230" can be parsed as offset 2:00 (only first digits are valid for | |
1670 | // extended format), but it can be parsed as offset 2:30 with basic format. We use longer result. | |
1671 | ParsePosition posBasic(start + 1); | |
1672 | int32_t tmpOffset = parseAbuttingAsciiOffsetFields(text, posBasic, FIELDS_H, FIELDS_HMS, FALSE); | |
1673 | if (posBasic.getErrorIndex() == -1 && posBasic.getIndex() > posOffset.getIndex()) { | |
1674 | offset = tmpOffset; | |
1675 | posOffset.setIndex(posBasic.getIndex()); | |
1676 | } | |
1677 | } | |
1678 | ||
1679 | if (posOffset.getErrorIndex() != -1) { | |
1680 | pos.setErrorIndex(start); | |
1681 | return 0; | |
1682 | } | |
1683 | ||
1684 | pos.setIndex(posOffset.getIndex()); | |
1685 | if (hasDigitOffset) { | |
1686 | *hasDigitOffset = TRUE; | |
1687 | } | |
1688 | return sign * offset; | |
1689 | } | |
1690 | ||
1691 | int32_t | |
51004dcb | 1692 | TimeZoneFormat::parseOffsetLocalizedGMT(const UnicodeString& text, ParsePosition& pos, UBool isShort, UBool* hasDigitOffset) const { |
4388f060 | 1693 | int32_t start = pos.getIndex(); |
4388f060 | 1694 | int32_t offset = 0; |
51004dcb | 1695 | int32_t parsedLength = 0; |
4388f060 A |
1696 | |
1697 | if (hasDigitOffset) { | |
1698 | *hasDigitOffset = FALSE; | |
1699 | } | |
1700 | ||
51004dcb | 1701 | offset = parseOffsetLocalizedGMTPattern(text, start, isShort, parsedLength); |
4388f060 | 1702 | |
51004dcb A |
1703 | // For now, parseOffsetLocalizedGMTPattern handles both long and short |
1704 | // formats, no matter isShort is true or false. This might be changed in future | |
1705 | // when strict parsing is necessary, or different set of patterns are used for | |
1706 | // short/long formats. | |
1707 | #if 0 | |
1708 | if (parsedLength == 0) { | |
1709 | offset = parseOffsetLocalizedGMTPattern(text, start, !isShort, parsedLength); | |
1710 | } | |
1711 | #endif | |
4388f060 | 1712 | |
51004dcb | 1713 | if (parsedLength > 0) { |
4388f060 A |
1714 | if (hasDigitOffset) { |
1715 | *hasDigitOffset = TRUE; | |
1716 | } | |
51004dcb | 1717 | pos.setIndex(start + parsedLength); |
4388f060 A |
1718 | return offset; |
1719 | } | |
1720 | ||
1721 | // Try the default patterns | |
4388f060 A |
1722 | offset = parseOffsetDefaultLocalizedGMT(text, start, parsedLength); |
1723 | if (parsedLength > 0) { | |
1724 | if (hasDigitOffset) { | |
1725 | *hasDigitOffset = TRUE; | |
1726 | } | |
1727 | pos.setIndex(start + parsedLength); | |
1728 | return offset; | |
1729 | } | |
1730 | ||
1731 | // Check if this is a GMT zero format | |
1732 | if (text.caseCompare(start, fGMTZeroFormat.length(), fGMTZeroFormat, 0) == 0) { | |
1733 | pos.setIndex(start + fGMTZeroFormat.length()); | |
1734 | return 0; | |
1735 | } | |
1736 | ||
1737 | // Check if this is a default GMT zero format | |
1738 | for (int32_t i = 0; ALT_GMT_STRINGS[i][0] != 0; i++) { | |
1739 | const UChar* defGMTZero = ALT_GMT_STRINGS[i]; | |
1740 | int32_t defGMTZeroLen = u_strlen(defGMTZero); | |
1741 | if (text.caseCompare(start, defGMTZeroLen, defGMTZero, 0) == 0) { | |
1742 | pos.setIndex(start + defGMTZeroLen); | |
1743 | return 0; | |
1744 | } | |
1745 | } | |
1746 | ||
1747 | // Nothing matched | |
1748 | pos.setErrorIndex(start); | |
1749 | return 0; | |
1750 | } | |
1751 | ||
1752 | int32_t | |
51004dcb A |
1753 | TimeZoneFormat::parseOffsetLocalizedGMTPattern(const UnicodeString& text, int32_t start, UBool /*isShort*/, int32_t& parsedLen) const { |
1754 | int32_t idx = start; | |
4388f060 | 1755 | int32_t offset = 0; |
51004dcb A |
1756 | UBool parsed = FALSE; |
1757 | ||
1758 | do { | |
1759 | // Prefix part | |
1760 | int32_t len = fGMTPatternPrefix.length(); | |
1761 | if (len > 0 && text.caseCompare(idx, len, fGMTPatternPrefix, 0) != 0) { | |
1762 | // prefix match failed | |
1763 | break; | |
1764 | } | |
1765 | idx += len; | |
1766 | ||
1767 | // Offset part | |
1768 | offset = parseOffsetFields(text, idx, FALSE, len); | |
1769 | if (len == 0) { | |
1770 | // offset field match failed | |
1771 | break; | |
1772 | } | |
1773 | idx += len; | |
1774 | ||
1775 | len = fGMTPatternSuffix.length(); | |
1776 | if (len > 0 && text.caseCompare(idx, len, fGMTPatternSuffix, 0) != 0) { | |
1777 | // no suffix match | |
1778 | break; | |
1779 | } | |
1780 | idx += len; | |
1781 | parsed = TRUE; | |
1782 | } while (FALSE); | |
1783 | ||
1784 | parsedLen = parsed ? idx - start : 0; | |
1785 | return offset; | |
1786 | } | |
1787 | ||
1788 | int32_t | |
1789 | TimeZoneFormat::parseOffsetFields(const UnicodeString& text, int32_t start, UBool /*isShort*/, int32_t& parsedLen) const { | |
1790 | int32_t outLen = 0; | |
1791 | int32_t offset = 0; | |
1792 | int32_t sign = 1; | |
4388f060 A |
1793 | |
1794 | parsedLen = 0; | |
1795 | ||
51004dcb A |
1796 | int32_t offsetH, offsetM, offsetS; |
1797 | offsetH = offsetM = offsetS = 0; | |
1798 | ||
4388f060 A |
1799 | for (int32_t patidx = 0; PARSE_GMT_OFFSET_TYPES[patidx] >= 0; patidx++) { |
1800 | int32_t gmtPatType = PARSE_GMT_OFFSET_TYPES[patidx]; | |
4388f060 A |
1801 | UVector* items = fGMTOffsetPatternItems[gmtPatType]; |
1802 | U_ASSERT(items != NULL); | |
4388f060 | 1803 | |
51004dcb A |
1804 | outLen = parseOffsetFieldsWithPattern(text, start, items, FALSE, offsetH, offsetM, offsetS); |
1805 | if (outLen > 0) { | |
1806 | sign = (gmtPatType == UTZFMT_PAT_POSITIVE_H || gmtPatType == UTZFMT_PAT_POSITIVE_HM || gmtPatType == UTZFMT_PAT_POSITIVE_HMS) ? | |
1807 | 1 : -1; | |
4388f060 A |
1808 | break; |
1809 | } | |
1810 | } | |
1811 | ||
51004dcb A |
1812 | if (outLen > 0 && fAbuttingOffsetHoursAndMinutes) { |
1813 | // When hours field is sabutting minutes field, | |
1814 | // the parse result above may not be appropriate. | |
1815 | // For example, "01020" is parsed as 01:02: above, | |
1816 | // but it should be parsed as 00:10:20. | |
1817 | int32_t tmpLen = 0; | |
1818 | int32_t tmpSign = 1; | |
0f5d89e8 A |
1819 | int32_t tmpH = 0; |
1820 | int32_t tmpM = 0; | |
1821 | int32_t tmpS = 0; | |
51004dcb A |
1822 | |
1823 | for (int32_t patidx = 0; PARSE_GMT_OFFSET_TYPES[patidx] >= 0; patidx++) { | |
1824 | int32_t gmtPatType = PARSE_GMT_OFFSET_TYPES[patidx]; | |
1825 | UVector* items = fGMTOffsetPatternItems[gmtPatType]; | |
1826 | U_ASSERT(items != NULL); | |
4388f060 | 1827 | |
51004dcb A |
1828 | // forcing parse to use single hour digit |
1829 | tmpLen = parseOffsetFieldsWithPattern(text, start, items, TRUE, tmpH, tmpM, tmpS); | |
1830 | if (tmpLen > 0) { | |
1831 | tmpSign = (gmtPatType == UTZFMT_PAT_POSITIVE_H || gmtPatType == UTZFMT_PAT_POSITIVE_HM || gmtPatType == UTZFMT_PAT_POSITIVE_HMS) ? | |
1832 | 1 : -1; | |
1833 | break; | |
1834 | } | |
1835 | } | |
1836 | if (tmpLen > outLen) { | |
1837 | // Better parse result with single hour digit | |
1838 | outLen = tmpLen; | |
1839 | sign = tmpSign; | |
1840 | offsetH = tmpH; | |
1841 | offsetM = tmpM; | |
1842 | offsetS = tmpS; | |
1843 | } | |
1844 | } | |
1845 | ||
1846 | if (outLen > 0) { | |
1847 | offset = ((((offsetH * 60) + offsetM) * 60) + offsetS) * 1000 * sign; | |
1848 | parsedLen = outLen; | |
4388f060 A |
1849 | } |
1850 | ||
1851 | return offset; | |
1852 | } | |
1853 | ||
51004dcb A |
1854 | int32_t |
1855 | TimeZoneFormat::parseOffsetFieldsWithPattern(const UnicodeString& text, int32_t start, | |
1856 | UVector* patternItems, UBool forceSingleHourDigit, int32_t& hour, int32_t& min, int32_t& sec) const { | |
1857 | UBool failed = FALSE; | |
1858 | int32_t offsetH, offsetM, offsetS; | |
1859 | offsetH = offsetM = offsetS = 0; | |
1860 | int32_t idx = start; | |
1861 | ||
1862 | for (int32_t i = 0; i < patternItems->size(); i++) { | |
57a6839d | 1863 | int32_t len = 0; |
51004dcb A |
1864 | const GMTOffsetField* field = (const GMTOffsetField*)patternItems->elementAt(i); |
1865 | GMTOffsetField::FieldType fieldType = field->getType(); | |
1866 | if (fieldType == GMTOffsetField::TEXT) { | |
1867 | const UChar* patStr = field->getPatternText(); | |
1868 | len = u_strlen(patStr); | |
0f5d89e8 A |
1869 | if (i == 0) { |
1870 | // When TimeZoneFormat parse() is called from SimpleDateFormat, | |
1871 | // leading space characters might be truncated. If the first pattern text | |
1872 | // starts with such character (e.g. Bidi control), then we need to | |
1873 | // skip the leading space charcters. | |
1874 | if (idx < text.length() && !PatternProps::isWhiteSpace(text.char32At(idx))) { | |
1875 | while (len > 0) { | |
1876 | UChar32 ch; | |
1877 | int32_t chLen; | |
1878 | U16_GET(patStr, 0, 0, len, ch) | |
1879 | if (PatternProps::isWhiteSpace(ch)) { | |
1880 | chLen = U16_LENGTH(ch); | |
1881 | len -= chLen; | |
1882 | patStr += chLen; | |
1883 | } | |
1884 | else { | |
1885 | break; | |
1886 | } | |
1887 | } | |
1888 | } | |
1889 | } | |
51004dcb A |
1890 | if (text.caseCompare(idx, len, patStr, 0) != 0) { |
1891 | failed = TRUE; | |
1892 | break; | |
1893 | } | |
1894 | idx += len; | |
1895 | } else { | |
1896 | if (fieldType == GMTOffsetField::HOUR) { | |
1897 | uint8_t maxDigits = forceSingleHourDigit ? 1 : 2; | |
1898 | offsetH = parseOffsetFieldWithLocalizedDigits(text, idx, 1, maxDigits, 0, MAX_OFFSET_HOUR, len); | |
1899 | } else if (fieldType == GMTOffsetField::MINUTE) { | |
1900 | offsetM = parseOffsetFieldWithLocalizedDigits(text, idx, 2, 2, 0, MAX_OFFSET_MINUTE, len); | |
1901 | } else if (fieldType == GMTOffsetField::SECOND) { | |
1902 | offsetS = parseOffsetFieldWithLocalizedDigits(text, idx, 2, 2, 0, MAX_OFFSET_SECOND, len); | |
1903 | } | |
1904 | ||
1905 | if (len == 0) { | |
1906 | failed = TRUE; | |
1907 | break; | |
1908 | } | |
1909 | idx += len; | |
1910 | } | |
1911 | } | |
1912 | ||
1913 | if (failed) { | |
1914 | hour = min = sec = 0; | |
1915 | return 0; | |
1916 | } | |
1917 | ||
1918 | hour = offsetH; | |
1919 | min = offsetM; | |
1920 | sec = offsetS; | |
1921 | ||
1922 | return idx - start; | |
1923 | } | |
1924 | ||
4388f060 A |
1925 | int32_t |
1926 | TimeZoneFormat::parseAbuttingOffsetFields(const UnicodeString& text, int32_t start, int32_t& parsedLen) const { | |
1927 | int32_t digits[MAX_OFFSET_DIGITS]; | |
1928 | int32_t parsed[MAX_OFFSET_DIGITS]; // accumulative offsets | |
1929 | ||
1930 | // Parse digits into int[] | |
1931 | int32_t idx = start; | |
1932 | int32_t len = 0; | |
1933 | int32_t numDigits = 0; | |
1934 | for (int32_t i = 0; i < MAX_OFFSET_DIGITS; i++) { | |
1935 | digits[i] = parseSingleLocalizedDigit(text, idx, len); | |
1936 | if (digits[i] < 0) { | |
1937 | break; | |
1938 | } | |
1939 | idx += len; | |
1940 | parsed[i] = idx - start; | |
1941 | numDigits++; | |
1942 | } | |
1943 | ||
1944 | if (numDigits == 0) { | |
1945 | parsedLen = 0; | |
1946 | return 0; | |
1947 | } | |
1948 | ||
1949 | int32_t offset = 0; | |
1950 | while (numDigits > 0) { | |
1951 | int32_t hour = 0; | |
1952 | int32_t min = 0; | |
1953 | int32_t sec = 0; | |
1954 | ||
1955 | U_ASSERT(numDigits > 0 && numDigits <= MAX_OFFSET_DIGITS); | |
1956 | switch (numDigits) { | |
1957 | case 1: // H | |
1958 | hour = digits[0]; | |
1959 | break; | |
1960 | case 2: // HH | |
1961 | hour = digits[0] * 10 + digits[1]; | |
1962 | break; | |
1963 | case 3: // Hmm | |
1964 | hour = digits[0]; | |
1965 | min = digits[1] * 10 + digits[2]; | |
1966 | break; | |
1967 | case 4: // HHmm | |
1968 | hour = digits[0] * 10 + digits[1]; | |
1969 | min = digits[2] * 10 + digits[3]; | |
1970 | break; | |
1971 | case 5: // Hmmss | |
1972 | hour = digits[0]; | |
1973 | min = digits[1] * 10 + digits[2]; | |
1974 | sec = digits[3] * 10 + digits[4]; | |
1975 | break; | |
1976 | case 6: // HHmmss | |
1977 | hour = digits[0] * 10 + digits[1]; | |
1978 | min = digits[2] * 10 + digits[3]; | |
1979 | sec = digits[4] * 10 + digits[5]; | |
1980 | break; | |
1981 | } | |
1982 | if (hour <= MAX_OFFSET_HOUR && min <= MAX_OFFSET_MINUTE && sec <= MAX_OFFSET_SECOND) { | |
1983 | // found a valid combination | |
1984 | offset = hour * MILLIS_PER_HOUR + min * MILLIS_PER_MINUTE + sec * MILLIS_PER_SECOND; | |
1985 | parsedLen = parsed[numDigits - 1]; | |
1986 | break; | |
1987 | } | |
1988 | numDigits--; | |
1989 | } | |
1990 | return offset; | |
1991 | } | |
1992 | ||
1993 | int32_t | |
1994 | TimeZoneFormat::parseOffsetDefaultLocalizedGMT(const UnicodeString& text, int start, int32_t& parsedLen) const { | |
1995 | int32_t idx = start; | |
1996 | int32_t offset = 0; | |
1997 | int32_t parsed = 0; | |
1998 | ||
1999 | do { | |
2000 | // check global default GMT alternatives | |
2001 | int32_t gmtLen = 0; | |
2002 | ||
2003 | for (int32_t i = 0; ALT_GMT_STRINGS[i][0] != 0; i++) { | |
2004 | const UChar* gmt = ALT_GMT_STRINGS[i]; | |
2005 | int32_t len = u_strlen(gmt); | |
2006 | if (text.caseCompare(start, len, gmt, 0) == 0) { | |
2007 | gmtLen = len; | |
2008 | break; | |
2009 | } | |
2010 | } | |
2011 | if (gmtLen == 0) { | |
2012 | break; | |
2013 | } | |
2014 | idx += gmtLen; | |
2015 | ||
2016 | // offset needs a sign char and a digit at minimum | |
2017 | if (idx + 1 >= text.length()) { | |
2018 | break; | |
2019 | } | |
2020 | ||
2021 | // parse sign | |
2022 | int32_t sign = 1; | |
2023 | UChar c = text.charAt(idx); | |
2024 | if (c == PLUS) { | |
2025 | sign = 1; | |
2026 | } else if (c == MINUS) { | |
2027 | sign = -1; | |
2028 | } else { | |
2029 | break; | |
2030 | } | |
2031 | idx++; | |
2032 | ||
2033 | // offset part | |
2034 | // try the default pattern with the separator first | |
2035 | int32_t lenWithSep = 0; | |
2036 | int32_t offsetWithSep = parseDefaultOffsetFields(text, idx, DEFAULT_GMT_OFFSET_SEP, lenWithSep); | |
2037 | if (lenWithSep == text.length() - idx) { | |
2038 | // maximum match | |
2039 | offset = offsetWithSep * sign; | |
2040 | idx += lenWithSep; | |
2041 | } else { | |
2042 | // try abutting field pattern | |
2043 | int32_t lenAbut = 0; | |
2044 | int32_t offsetAbut = parseAbuttingOffsetFields(text, idx, lenAbut); | |
2045 | ||
2046 | if (lenWithSep > lenAbut) { | |
2047 | offset = offsetWithSep * sign; | |
2048 | idx += lenWithSep; | |
2049 | } else { | |
2050 | offset = offsetAbut * sign; | |
2051 | idx += lenAbut; | |
2052 | } | |
2053 | } | |
2054 | parsed = idx - start; | |
2055 | } while (false); | |
2056 | ||
2057 | parsedLen = parsed; | |
2058 | return offset; | |
2059 | } | |
2060 | ||
2061 | int32_t | |
2062 | TimeZoneFormat::parseDefaultOffsetFields(const UnicodeString& text, int32_t start, UChar separator, int32_t& parsedLen) const { | |
2063 | int32_t max = text.length(); | |
2064 | int32_t idx = start; | |
2065 | int32_t len = 0; | |
2066 | int32_t hour = 0, min = 0, sec = 0; | |
2067 | ||
2068 | parsedLen = 0; | |
2069 | ||
2070 | do { | |
2071 | hour = parseOffsetFieldWithLocalizedDigits(text, idx, 1, 2, 0, MAX_OFFSET_HOUR, len); | |
2072 | if (len == 0) { | |
2073 | break; | |
2074 | } | |
2075 | idx += len; | |
2076 | ||
2077 | if (idx + 1 < max && text.charAt(idx) == separator) { | |
2078 | min = parseOffsetFieldWithLocalizedDigits(text, idx + 1, 2, 2, 0, MAX_OFFSET_MINUTE, len); | |
2079 | if (len == 0) { | |
2080 | break; | |
2081 | } | |
2082 | idx += (1 + len); | |
2083 | ||
2084 | if (idx + 1 < max && text.charAt(idx) == separator) { | |
2085 | sec = parseOffsetFieldWithLocalizedDigits(text, idx + 1, 2, 2, 0, MAX_OFFSET_SECOND, len); | |
2086 | if (len == 0) { | |
2087 | break; | |
2088 | } | |
2089 | idx += (1 + len); | |
2090 | } | |
2091 | } | |
2092 | } while (FALSE); | |
2093 | ||
2094 | if (idx == start) { | |
2095 | return 0; | |
2096 | } | |
2097 | ||
2098 | parsedLen = idx - start; | |
2099 | return hour * MILLIS_PER_HOUR + min * MILLIS_PER_MINUTE + sec * MILLIS_PER_SECOND; | |
2100 | } | |
2101 | ||
2102 | int32_t | |
2103 | TimeZoneFormat::parseOffsetFieldWithLocalizedDigits(const UnicodeString& text, int32_t start, uint8_t minDigits, uint8_t maxDigits, uint16_t minVal, uint16_t maxVal, int32_t& parsedLen) const { | |
2104 | parsedLen = 0; | |
2105 | ||
2106 | int32_t decVal = 0; | |
2107 | int32_t numDigits = 0; | |
2108 | int32_t idx = start; | |
2109 | int32_t digitLen = 0; | |
2110 | ||
2111 | while (idx < text.length() && numDigits < maxDigits) { | |
2112 | int32_t digit = parseSingleLocalizedDigit(text, idx, digitLen); | |
2113 | if (digit < 0) { | |
2114 | break; | |
2115 | } | |
2116 | int32_t tmpVal = decVal * 10 + digit; | |
2117 | if (tmpVal > maxVal) { | |
2118 | break; | |
2119 | } | |
2120 | decVal = tmpVal; | |
2121 | numDigits++; | |
2122 | idx += digitLen; | |
2123 | } | |
2124 | ||
2125 | // Note: maxVal is checked in the while loop | |
2126 | if (numDigits < minDigits || decVal < minVal) { | |
2127 | decVal = -1; | |
2128 | numDigits = 0; | |
2129 | } else { | |
2130 | parsedLen = idx - start; | |
2131 | } | |
2132 | ||
2133 | return decVal; | |
2134 | } | |
2135 | ||
2136 | int32_t | |
2137 | TimeZoneFormat::parseSingleLocalizedDigit(const UnicodeString& text, int32_t start, int32_t& len) const { | |
2138 | int32_t digit = -1; | |
2139 | len = 0; | |
2140 | if (start < text.length()) { | |
2141 | UChar32 cp = text.char32At(start); | |
2142 | ||
2143 | // First, try digits configured for this instance | |
2144 | for (int32_t i = 0; i < 10; i++) { | |
2145 | if (cp == fGMTOffsetDigits[i]) { | |
2146 | digit = i; | |
2147 | break; | |
2148 | } | |
2149 | } | |
2150 | // If failed, check if this is a Unicode digit | |
2151 | if (digit < 0) { | |
2152 | int32_t tmp = u_charDigitValue(cp); | |
2153 | digit = (tmp >= 0 && tmp <= 9) ? tmp : -1; | |
2154 | } | |
2155 | ||
2156 | if (digit >= 0) { | |
2157 | int32_t next = text.moveIndex32(start, 1); | |
2158 | len = next - start; | |
2159 | } | |
2160 | } | |
2161 | return digit; | |
2162 | } | |
2163 | ||
2164 | UnicodeString& | |
2165 | TimeZoneFormat::formatOffsetWithAsciiDigits(int32_t offset, UChar sep, OffsetFields minFields, OffsetFields maxFields, UnicodeString& result) { | |
2166 | U_ASSERT(maxFields >= minFields); | |
2167 | U_ASSERT(offset > -MAX_OFFSET && offset < MAX_OFFSET); | |
2168 | ||
2169 | UChar sign = PLUS; | |
2170 | if (offset < 0) { | |
2171 | sign = MINUS; | |
2172 | offset = -offset; | |
2173 | } | |
2174 | result.setTo(sign); | |
2175 | ||
2176 | int fields[3]; | |
2177 | fields[0] = offset / MILLIS_PER_HOUR; | |
2178 | offset = offset % MILLIS_PER_HOUR; | |
2179 | fields[1] = offset / MILLIS_PER_MINUTE; | |
2180 | offset = offset % MILLIS_PER_MINUTE; | |
2181 | fields[2] = offset / MILLIS_PER_SECOND; | |
2182 | ||
2183 | U_ASSERT(fields[0] >= 0 && fields[0] <= MAX_OFFSET_HOUR); | |
2184 | U_ASSERT(fields[1] >= 0 && fields[1] <= MAX_OFFSET_MINUTE); | |
2185 | U_ASSERT(fields[2] >= 0 && fields[2] <= MAX_OFFSET_SECOND); | |
2186 | ||
2187 | int32_t lastIdx = maxFields; | |
2188 | while (lastIdx > minFields) { | |
2189 | if (fields[lastIdx] != 0) { | |
2190 | break; | |
2191 | } | |
2192 | lastIdx--; | |
2193 | } | |
2194 | ||
2195 | for (int32_t idx = 0; idx <= lastIdx; idx++) { | |
2196 | if (sep && idx != 0) { | |
2197 | result.append(sep); | |
2198 | } | |
2199 | result.append((UChar)(0x0030 + fields[idx]/10)); | |
2200 | result.append((UChar)(0x0030 + fields[idx]%10)); | |
2201 | } | |
2202 | ||
2203 | return result; | |
2204 | } | |
2205 | ||
2206 | int32_t | |
2207 | TimeZoneFormat::parseAbuttingAsciiOffsetFields(const UnicodeString& text, ParsePosition& pos, OffsetFields minFields, OffsetFields maxFields, UBool fixedHourWidth) { | |
2208 | int32_t start = pos.getIndex(); | |
2209 | ||
2210 | int32_t minDigits = 2 * (minFields + 1) - (fixedHourWidth ? 0 : 1); | |
2211 | int32_t maxDigits = 2 * (maxFields + 1); | |
2212 | ||
2213 | U_ASSERT(maxDigits <= MAX_OFFSET_DIGITS); | |
2214 | ||
51004dcb | 2215 | int32_t digits[MAX_OFFSET_DIGITS] = {}; |
4388f060 A |
2216 | int32_t numDigits = 0; |
2217 | int32_t idx = start; | |
2218 | while (numDigits < maxDigits && idx < text.length()) { | |
2219 | UChar uch = text.charAt(idx); | |
2220 | int32_t digit = DIGIT_VAL(uch); | |
2221 | if (digit < 0) { | |
2222 | break; | |
2223 | } | |
2224 | digits[numDigits] = digit; | |
2225 | numDigits++; | |
2226 | idx++; | |
2227 | } | |
2228 | ||
2229 | if (fixedHourWidth && (numDigits & 1)) { | |
2230 | // Fixed digits, so the number of digits must be even number. Truncating. | |
2231 | numDigits--; | |
2232 | } | |
2233 | ||
2234 | if (numDigits < minDigits) { | |
2235 | pos.setErrorIndex(start); | |
2236 | return 0; | |
2237 | } | |
2238 | ||
2239 | int32_t hour = 0, min = 0, sec = 0; | |
2240 | UBool bParsed = FALSE; | |
2241 | while (numDigits >= minDigits) { | |
2242 | switch (numDigits) { | |
2243 | case 1: //H | |
2244 | hour = digits[0]; | |
2245 | break; | |
2246 | case 2: //HH | |
2247 | hour = digits[0] * 10 + digits[1]; | |
2248 | break; | |
2249 | case 3: //Hmm | |
2250 | hour = digits[0]; | |
2251 | min = digits[1] * 10 + digits[2]; | |
2252 | break; | |
2253 | case 4: //HHmm | |
2254 | hour = digits[0] * 10 + digits[1]; | |
2255 | min = digits[2] * 10 + digits[3]; | |
2256 | break; | |
2257 | case 5: //Hmmss | |
2258 | hour = digits[0]; | |
2259 | min = digits[1] * 10 + digits[2]; | |
2260 | sec = digits[3] * 10 + digits[4]; | |
2261 | break; | |
2262 | case 6: //HHmmss | |
2263 | hour = digits[0] * 10 + digits[1]; | |
2264 | min = digits[2] * 10 + digits[3]; | |
2265 | sec = digits[4] * 10 + digits[5]; | |
2266 | break; | |
2267 | } | |
2268 | ||
2269 | if (hour <= MAX_OFFSET_HOUR && min <= MAX_OFFSET_MINUTE && sec <= MAX_OFFSET_SECOND) { | |
2270 | // Successfully parsed | |
2271 | bParsed = true; | |
2272 | break; | |
2273 | } | |
2274 | ||
2275 | // Truncating | |
2276 | numDigits -= (fixedHourWidth ? 2 : 1); | |
2277 | hour = min = sec = 0; | |
2278 | } | |
2279 | ||
2280 | if (!bParsed) { | |
2281 | pos.setErrorIndex(start); | |
2282 | return 0; | |
2283 | } | |
2284 | pos.setIndex(start + numDigits); | |
2285 | return ((((hour * 60) + min) * 60) + sec) * 1000; | |
2286 | } | |
2287 | ||
2288 | int32_t | |
51004dcb | 2289 | TimeZoneFormat::parseAsciiOffsetFields(const UnicodeString& text, ParsePosition& pos, UChar sep, OffsetFields minFields, OffsetFields maxFields) { |
4388f060 A |
2290 | int32_t start = pos.getIndex(); |
2291 | int32_t fieldVal[] = {0, 0, 0}; | |
2292 | int32_t fieldLen[] = {0, -1, -1}; | |
2293 | for (int32_t idx = start, fieldIdx = 0; idx < text.length() && fieldIdx <= maxFields; idx++) { | |
2294 | UChar c = text.charAt(idx); | |
2295 | if (c == sep) { | |
51004dcb A |
2296 | if (fieldIdx == 0) { |
2297 | if (fieldLen[0] == 0) { | |
2298 | // no hours field | |
2299 | break; | |
2300 | } | |
4388f060 | 2301 | // 1 digit hour, move to next field |
4388f060 | 2302 | } else { |
51004dcb A |
2303 | if (fieldLen[fieldIdx] != -1) { |
2304 | // premature minute or seconds field | |
2305 | break; | |
2306 | } | |
2307 | fieldLen[fieldIdx] = 0; | |
4388f060 A |
2308 | } |
2309 | continue; | |
51004dcb A |
2310 | } else if (fieldLen[fieldIdx] == -1) { |
2311 | // no separator after 2 digit field | |
2312 | break; | |
4388f060 A |
2313 | } |
2314 | int32_t digit = DIGIT_VAL(c); | |
2315 | if (digit < 0) { | |
2316 | // not a digit | |
2317 | break; | |
2318 | } | |
2319 | fieldVal[fieldIdx] = fieldVal[fieldIdx] * 10 + digit; | |
2320 | fieldLen[fieldIdx]++; | |
2321 | if (fieldLen[fieldIdx] >= 2) { | |
2322 | // parsed 2 digits, move to next field | |
2323 | fieldIdx++; | |
2324 | } | |
2325 | } | |
2326 | ||
2327 | int32_t offset = 0; | |
2328 | int32_t parsedLen = 0; | |
2329 | int32_t parsedFields = -1; | |
2330 | do { | |
2331 | // hour | |
51004dcb | 2332 | if (fieldLen[0] == 0) { |
4388f060 A |
2333 | break; |
2334 | } | |
2335 | if (fieldVal[0] > MAX_OFFSET_HOUR) { | |
4388f060 A |
2336 | offset = (fieldVal[0] / 10) * MILLIS_PER_HOUR; |
2337 | parsedFields = FIELDS_H; | |
2338 | parsedLen = 1; | |
2339 | break; | |
2340 | } | |
2341 | offset = fieldVal[0] * MILLIS_PER_HOUR; | |
2342 | parsedLen = fieldLen[0]; | |
2343 | parsedFields = FIELDS_H; | |
2344 | ||
2345 | // minute | |
2346 | if (fieldLen[1] != 2 || fieldVal[1] > MAX_OFFSET_MINUTE) { | |
2347 | break; | |
2348 | } | |
2349 | offset += fieldVal[1] * MILLIS_PER_MINUTE; | |
2350 | parsedLen += (1 + fieldLen[1]); | |
2351 | parsedFields = FIELDS_HM; | |
2352 | ||
2353 | // second | |
2354 | if (fieldLen[2] != 2 || fieldVal[2] > MAX_OFFSET_SECOND) { | |
2355 | break; | |
2356 | } | |
2357 | offset += fieldVal[2] * MILLIS_PER_SECOND; | |
2358 | parsedLen += (1 + fieldLen[2]); | |
2359 | parsedFields = FIELDS_HMS; | |
2360 | } while (false); | |
2361 | ||
2362 | if (parsedFields < minFields) { | |
2363 | pos.setErrorIndex(start); | |
2364 | return 0; | |
2365 | } | |
2366 | ||
2367 | pos.setIndex(start + parsedLen); | |
2368 | return offset; | |
2369 | } | |
2370 | ||
2371 | void | |
2372 | TimeZoneFormat::appendOffsetDigits(UnicodeString& buf, int32_t n, uint8_t minDigits) const { | |
2373 | U_ASSERT(n >= 0 && n < 60); | |
2374 | int32_t numDigits = n >= 10 ? 2 : 1; | |
2375 | for (int32_t i = 0; i < minDigits - numDigits; i++) { | |
2376 | buf.append(fGMTOffsetDigits[0]); | |
2377 | } | |
2378 | if (numDigits == 2) { | |
2379 | buf.append(fGMTOffsetDigits[n / 10]); | |
2380 | } | |
2381 | buf.append(fGMTOffsetDigits[n % 10]); | |
2382 | } | |
2383 | ||
2384 | // ------------------------------------------------------------------ | |
2385 | // Private misc | |
2386 | void | |
2387 | TimeZoneFormat::initGMTPattern(const UnicodeString& gmtPattern, UErrorCode& status) { | |
2388 | if (U_FAILURE(status)) { | |
2389 | return; | |
2390 | } | |
2391 | // This implementation not perfect, but sufficient practically. | |
2392 | int32_t idx = gmtPattern.indexOf(ARG0, ARG0_LEN, 0); | |
2393 | if (idx < 0) { | |
2394 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
2395 | return; | |
2396 | } | |
2397 | fGMTPattern.setTo(gmtPattern); | |
2398 | unquote(gmtPattern.tempSubString(0, idx), fGMTPatternPrefix); | |
2399 | unquote(gmtPattern.tempSubString(idx + ARG0_LEN), fGMTPatternSuffix); | |
2400 | } | |
2401 | ||
2402 | UnicodeString& | |
2403 | TimeZoneFormat::unquote(const UnicodeString& pattern, UnicodeString& result) { | |
2404 | if (pattern.indexOf(SINGLEQUOTE) < 0) { | |
2405 | result.setTo(pattern); | |
2406 | return result; | |
2407 | } | |
2408 | result.remove(); | |
2409 | UBool isPrevQuote = FALSE; | |
2410 | UBool inQuote = FALSE; | |
2411 | for (int32_t i = 0; i < pattern.length(); i++) { | |
2412 | UChar c = pattern.charAt(i); | |
2413 | if (c == SINGLEQUOTE) { | |
2414 | if (isPrevQuote) { | |
2415 | result.append(c); | |
2416 | isPrevQuote = FALSE; | |
2417 | } else { | |
2418 | isPrevQuote = TRUE; | |
2419 | } | |
2420 | inQuote = !inQuote; | |
2421 | } else { | |
2422 | isPrevQuote = FALSE; | |
2423 | result.append(c); | |
2424 | } | |
2425 | } | |
2426 | return result; | |
2427 | } | |
2428 | ||
2429 | UVector* | |
2430 | TimeZoneFormat::parseOffsetPattern(const UnicodeString& pattern, OffsetFields required, UErrorCode& status) { | |
2431 | if (U_FAILURE(status)) { | |
2432 | return NULL; | |
2433 | } | |
2434 | UVector* result = new UVector(deleteGMTOffsetField, NULL, status); | |
2435 | if (result == NULL) { | |
2436 | status = U_MEMORY_ALLOCATION_ERROR; | |
2437 | return NULL; | |
2438 | } | |
2439 | ||
2440 | int32_t checkBits = 0; | |
2441 | UBool isPrevQuote = FALSE; | |
2442 | UBool inQuote = FALSE; | |
b331163b A |
2443 | UChar textBuf[32]; |
2444 | UnicodeString text(textBuf, 0, UPRV_LENGTHOF(textBuf)); | |
4388f060 A |
2445 | GMTOffsetField::FieldType itemType = GMTOffsetField::TEXT; |
2446 | int32_t itemLength = 1; | |
2447 | ||
2448 | for (int32_t i = 0; i < pattern.length(); i++) { | |
2449 | UChar ch = pattern.charAt(i); | |
2450 | if (ch == SINGLEQUOTE) { | |
2451 | if (isPrevQuote) { | |
2452 | text.append(SINGLEQUOTE); | |
2453 | isPrevQuote = FALSE; | |
2454 | } else { | |
2455 | isPrevQuote = TRUE; | |
2456 | if (itemType != GMTOffsetField::TEXT) { | |
2457 | if (GMTOffsetField::isValid(itemType, itemLength)) { | |
0f5d89e8 | 2458 | GMTOffsetField* fld = GMTOffsetField::createTimeField(itemType, static_cast<uint8_t>(itemLength), status); |
4388f060 A |
2459 | result->addElement(fld, status); |
2460 | if (U_FAILURE(status)) { | |
2461 | break; | |
2462 | } | |
2463 | } else { | |
2464 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
2465 | break; | |
2466 | } | |
2467 | itemType = GMTOffsetField::TEXT; | |
2468 | } | |
2469 | } | |
2470 | inQuote = !inQuote; | |
2471 | } else { | |
2472 | isPrevQuote = FALSE; | |
2473 | if (inQuote) { | |
2474 | text.append(ch); | |
2475 | } else { | |
2476 | GMTOffsetField::FieldType tmpType = GMTOffsetField::getTypeByLetter(ch); | |
2477 | if (tmpType != GMTOffsetField::TEXT) { | |
2478 | // an offset time pattern character | |
2479 | if (tmpType == itemType) { | |
2480 | itemLength++; | |
2481 | } else { | |
2482 | if (itemType == GMTOffsetField::TEXT) { | |
2483 | if (text.length() > 0) { | |
2484 | GMTOffsetField* textfld = GMTOffsetField::createText(text, status); | |
2485 | result->addElement(textfld, status); | |
2486 | if (U_FAILURE(status)) { | |
2487 | break; | |
2488 | } | |
2489 | text.remove(); | |
2490 | } | |
2491 | } else { | |
2492 | if (GMTOffsetField::isValid(itemType, itemLength)) { | |
0f5d89e8 | 2493 | GMTOffsetField* fld = GMTOffsetField::createTimeField(itemType, static_cast<uint8_t>(itemLength), status); |
4388f060 A |
2494 | result->addElement(fld, status); |
2495 | if (U_FAILURE(status)) { | |
2496 | break; | |
2497 | } | |
2498 | } else { | |
2499 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
2500 | break; | |
2501 | } | |
2502 | } | |
2503 | itemType = tmpType; | |
2504 | itemLength = 1; | |
2505 | checkBits |= tmpType; | |
2506 | } | |
2507 | } else { | |
2508 | // a string literal | |
2509 | if (itemType != GMTOffsetField::TEXT) { | |
2510 | if (GMTOffsetField::isValid(itemType, itemLength)) { | |
0f5d89e8 | 2511 | GMTOffsetField* fld = GMTOffsetField::createTimeField(itemType, static_cast<uint8_t>(itemLength), status); |
4388f060 A |
2512 | result->addElement(fld, status); |
2513 | if (U_FAILURE(status)) { | |
2514 | break; | |
2515 | } | |
2516 | } else { | |
2517 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
2518 | break; | |
2519 | } | |
2520 | itemType = GMTOffsetField::TEXT; | |
2521 | } | |
2522 | text.append(ch); | |
2523 | } | |
2524 | } | |
2525 | } | |
2526 | } | |
2527 | // handle last item | |
2528 | if (U_SUCCESS(status)) { | |
2529 | if (itemType == GMTOffsetField::TEXT) { | |
2530 | if (text.length() > 0) { | |
2531 | GMTOffsetField* tfld = GMTOffsetField::createText(text, status); | |
2532 | result->addElement(tfld, status); | |
2533 | } | |
2534 | } else { | |
2535 | if (GMTOffsetField::isValid(itemType, itemLength)) { | |
0f5d89e8 | 2536 | GMTOffsetField* fld = GMTOffsetField::createTimeField(itemType, static_cast<uint8_t>(itemLength), status); |
4388f060 A |
2537 | result->addElement(fld, status); |
2538 | } else { | |
2539 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
2540 | } | |
2541 | } | |
2542 | ||
2543 | // Check all required fields are set | |
2544 | if (U_SUCCESS(status)) { | |
2545 | int32_t reqBits = 0; | |
2546 | switch (required) { | |
2547 | case FIELDS_H: | |
2548 | reqBits = GMTOffsetField::HOUR; | |
2549 | break; | |
2550 | case FIELDS_HM: | |
2551 | reqBits = GMTOffsetField::HOUR | GMTOffsetField::MINUTE; | |
2552 | break; | |
2553 | case FIELDS_HMS: | |
2554 | reqBits = GMTOffsetField::HOUR | GMTOffsetField::MINUTE | GMTOffsetField::SECOND; | |
2555 | break; | |
2556 | } | |
2557 | if (checkBits == reqBits) { | |
2558 | // all required fields are set, no extra fields | |
2559 | return result; | |
2560 | } | |
2561 | } | |
2562 | } | |
2563 | ||
2564 | // error | |
2565 | delete result; | |
2566 | return NULL; | |
2567 | } | |
2568 | ||
2569 | UnicodeString& | |
51004dcb A |
2570 | TimeZoneFormat::expandOffsetPattern(const UnicodeString& offsetHM, UnicodeString& result, UErrorCode& status) { |
2571 | result.setToBogus(); | |
2572 | if (U_FAILURE(status)) { | |
2573 | return result; | |
2574 | } | |
4388f060 A |
2575 | U_ASSERT(u_strlen(DEFAULT_GMT_OFFSET_MINUTE_PATTERN) == 2); |
2576 | ||
2577 | int32_t idx_mm = offsetHM.indexOf(DEFAULT_GMT_OFFSET_MINUTE_PATTERN, 2, 0); | |
2578 | if (idx_mm < 0) { | |
51004dcb A |
2579 | // Bad time zone hour pattern data |
2580 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
4388f060 A |
2581 | return result; |
2582 | } | |
2583 | ||
2584 | UnicodeString sep; | |
51004dcb | 2585 | int32_t idx_H = offsetHM.tempSubString(0, idx_mm).lastIndexOf((UChar)0x0048 /* H */); |
4388f060 A |
2586 | if (idx_H >= 0) { |
2587 | sep = offsetHM.tempSubString(idx_H + 1, idx_mm - (idx_H + 1)); | |
2588 | } | |
2589 | result.setTo(offsetHM.tempSubString(0, idx_mm + 2)); | |
2590 | result.append(sep); | |
2591 | result.append(DEFAULT_GMT_OFFSET_SECOND_PATTERN, -1); | |
2592 | result.append(offsetHM.tempSubString(idx_mm + 2)); | |
2593 | return result; | |
2594 | } | |
2595 | ||
51004dcb A |
2596 | UnicodeString& |
2597 | TimeZoneFormat::truncateOffsetPattern(const UnicodeString& offsetHM, UnicodeString& result, UErrorCode& status) { | |
2598 | result.setToBogus(); | |
2599 | if (U_FAILURE(status)) { | |
2600 | return result; | |
2601 | } | |
2602 | U_ASSERT(u_strlen(DEFAULT_GMT_OFFSET_MINUTE_PATTERN) == 2); | |
2603 | ||
2604 | int32_t idx_mm = offsetHM.indexOf(DEFAULT_GMT_OFFSET_MINUTE_PATTERN, 2, 0); | |
2605 | if (idx_mm < 0) { | |
2606 | // Bad time zone hour pattern data | |
2607 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
2608 | return result; | |
2609 | } | |
2610 | UChar HH[] = {0x0048, 0x0048}; | |
2611 | int32_t idx_HH = offsetHM.tempSubString(0, idx_mm).lastIndexOf(HH, 2, 0); | |
2612 | if (idx_HH >= 0) { | |
2613 | return result.setTo(offsetHM.tempSubString(0, idx_HH + 2)); | |
2614 | } | |
2615 | int32_t idx_H = offsetHM.tempSubString(0, idx_mm).lastIndexOf((UChar)0x0048, 0); | |
2616 | if (idx_H >= 0) { | |
2617 | return result.setTo(offsetHM.tempSubString(0, idx_H + 1)); | |
2618 | } | |
2619 | // Bad time zone hour pattern data | |
2620 | status = U_ILLEGAL_ARGUMENT_ERROR; | |
2621 | return result; | |
2622 | } | |
2623 | ||
4388f060 A |
2624 | void |
2625 | TimeZoneFormat::initGMTOffsetPatterns(UErrorCode& status) { | |
51004dcb | 2626 | for (int32_t type = 0; type < UTZFMT_PAT_COUNT; type++) { |
4388f060 | 2627 | switch (type) { |
51004dcb A |
2628 | case UTZFMT_PAT_POSITIVE_H: |
2629 | case UTZFMT_PAT_NEGATIVE_H: | |
2630 | fGMTOffsetPatternItems[type] = parseOffsetPattern(fGMTOffsetPatterns[type], FIELDS_H, status); | |
2631 | break; | |
4388f060 A |
2632 | case UTZFMT_PAT_POSITIVE_HM: |
2633 | case UTZFMT_PAT_NEGATIVE_HM: | |
2634 | fGMTOffsetPatternItems[type] = parseOffsetPattern(fGMTOffsetPatterns[type], FIELDS_HM, status); | |
2635 | break; | |
2636 | case UTZFMT_PAT_POSITIVE_HMS: | |
2637 | case UTZFMT_PAT_NEGATIVE_HMS: | |
2638 | fGMTOffsetPatternItems[type] = parseOffsetPattern(fGMTOffsetPatterns[type], FIELDS_HMS, status); | |
2639 | break; | |
2640 | } | |
2641 | } | |
51004dcb A |
2642 | checkAbuttingHoursAndMinutes(); |
2643 | } | |
2644 | ||
2645 | void | |
2646 | TimeZoneFormat::checkAbuttingHoursAndMinutes() { | |
2647 | fAbuttingOffsetHoursAndMinutes= FALSE; | |
2648 | for (int32_t type = 0; type < UTZFMT_PAT_COUNT; type++) { | |
2649 | UBool afterH = FALSE; | |
2650 | UVector *items = fGMTOffsetPatternItems[type]; | |
2651 | for (int32_t i = 0; i < items->size(); i++) { | |
2652 | const GMTOffsetField* item = (GMTOffsetField*)items->elementAt(i); | |
3d1f044b A |
2653 | GMTOffsetField::FieldType fieldType = item->getType(); |
2654 | if (fieldType != GMTOffsetField::TEXT) { | |
51004dcb A |
2655 | if (afterH) { |
2656 | fAbuttingOffsetHoursAndMinutes = TRUE; | |
2657 | break; | |
3d1f044b | 2658 | } else if (fieldType == GMTOffsetField::HOUR) { |
51004dcb A |
2659 | afterH = TRUE; |
2660 | } | |
2661 | } else if (afterH) { | |
2662 | break; | |
2663 | } | |
2664 | } | |
2665 | if (fAbuttingOffsetHoursAndMinutes) { | |
2666 | break; | |
2667 | } | |
2668 | } | |
4388f060 A |
2669 | } |
2670 | ||
2671 | UBool | |
2672 | TimeZoneFormat::toCodePoints(const UnicodeString& str, UChar32* codeArray, int32_t size) { | |
2673 | int32_t count = str.countChar32(); | |
2674 | if (count != size) { | |
2675 | return FALSE; | |
2676 | } | |
2677 | ||
2678 | for (int32_t idx = 0, start = 0; idx < size; idx++) { | |
2679 | codeArray[idx] = str.char32At(start); | |
2680 | start = str.moveIndex32(start, 1); | |
2681 | } | |
2682 | ||
2683 | return TRUE; | |
2684 | } | |
2685 | ||
2686 | TimeZone* | |
2687 | TimeZoneFormat::createTimeZoneForOffset(int32_t offset) const { | |
2688 | if (offset == 0) { | |
2689 | // when offset is 0, we should use "Etc/GMT" | |
b331163b | 2690 | return TimeZone::createTimeZone(UnicodeString(TRUE, TZID_GMT, -1)); |
4388f060 A |
2691 | } |
2692 | return ZoneMeta::createCustomTimeZone(offset); | |
2693 | } | |
2694 | ||
2695 | UTimeZoneFormatTimeType | |
2696 | TimeZoneFormat::getTimeType(UTimeZoneNameType nameType) { | |
2697 | switch (nameType) { | |
2698 | case UTZNM_LONG_STANDARD: | |
2699 | case UTZNM_SHORT_STANDARD: | |
2700 | return UTZFMT_TIME_TYPE_STANDARD; | |
2701 | ||
2702 | case UTZNM_LONG_DAYLIGHT: | |
2703 | case UTZNM_SHORT_DAYLIGHT: | |
2704 | return UTZFMT_TIME_TYPE_DAYLIGHT; | |
2705 | ||
2706 | default: | |
b331163b | 2707 | return UTZFMT_TIME_TYPE_UNKNOWN; |
4388f060 | 2708 | } |
4388f060 A |
2709 | } |
2710 | ||
2711 | UnicodeString& | |
2712 | TimeZoneFormat::getTimeZoneID(const TimeZoneNames::MatchInfoCollection* matches, int32_t idx, UnicodeString& tzID) const { | |
2713 | if (!matches->getTimeZoneIDAt(idx, tzID)) { | |
b331163b A |
2714 | UChar mzIDBuf[32]; |
2715 | UnicodeString mzID(mzIDBuf, 0, UPRV_LENGTHOF(mzIDBuf)); | |
4388f060 A |
2716 | if (matches->getMetaZoneIDAt(idx, mzID)) { |
2717 | fTimeZoneNames->getReferenceZoneID(mzID, fTargetRegion, tzID); | |
2718 | } | |
2719 | } | |
2720 | return tzID; | |
2721 | } | |
2722 | ||
51004dcb A |
2723 | |
2724 | class ZoneIdMatchHandler : public TextTrieMapSearchResultHandler { | |
2725 | public: | |
2726 | ZoneIdMatchHandler(); | |
2727 | virtual ~ZoneIdMatchHandler(); | |
2728 | ||
2729 | UBool handleMatch(int32_t matchLength, const CharacterNode *node, UErrorCode &status); | |
2730 | const UChar* getID(); | |
2731 | int32_t getMatchLen(); | |
2732 | private: | |
2733 | int32_t fLen; | |
2734 | const UChar* fID; | |
2735 | }; | |
2736 | ||
2737 | ZoneIdMatchHandler::ZoneIdMatchHandler() | |
2738 | : fLen(0), fID(NULL) { | |
2739 | } | |
2740 | ||
2741 | ZoneIdMatchHandler::~ZoneIdMatchHandler() { | |
2742 | } | |
2743 | ||
2744 | UBool | |
2745 | ZoneIdMatchHandler::handleMatch(int32_t matchLength, const CharacterNode *node, UErrorCode &status) { | |
2746 | if (U_FAILURE(status)) { | |
2747 | return FALSE; | |
2748 | } | |
2749 | if (node->hasValues()) { | |
2750 | const UChar* id = (const UChar*)node->getValue(0); | |
2751 | if (id != NULL) { | |
2752 | if (fLen < matchLength) { | |
2753 | fID = id; | |
2754 | fLen = matchLength; | |
2755 | } | |
2756 | } | |
2757 | } | |
2758 | return TRUE; | |
2759 | } | |
2760 | ||
2761 | const UChar* | |
2762 | ZoneIdMatchHandler::getID() { | |
2763 | return fID; | |
2764 | } | |
2765 | ||
2766 | int32_t | |
2767 | ZoneIdMatchHandler::getMatchLen() { | |
2768 | return fLen; | |
2769 | } | |
2770 | ||
57a6839d A |
2771 | |
2772 | static void U_CALLCONV initZoneIdTrie(UErrorCode &status) { | |
2773 | U_ASSERT(gZoneIdTrie == NULL); | |
2774 | ucln_i18n_registerCleanup(UCLN_I18N_TIMEZONEFORMAT, tzfmt_cleanup); | |
2775 | gZoneIdTrie = new TextTrieMap(TRUE, NULL); // No deleter, because values are pooled by ZoneMeta | |
2776 | if (gZoneIdTrie == NULL) { | |
2777 | status = U_MEMORY_ALLOCATION_ERROR; | |
2778 | return; | |
2779 | } | |
2780 | StringEnumeration *tzenum = TimeZone::createEnumeration(); | |
2781 | const UnicodeString *id; | |
0f5d89e8 | 2782 | while ((id = tzenum->snext(status)) != NULL) { |
57a6839d A |
2783 | const UChar* uid = ZoneMeta::findTimeZoneID(*id); |
2784 | if (uid) { | |
2785 | gZoneIdTrie->put(uid, const_cast<UChar *>(uid), status); | |
2786 | } | |
2787 | } | |
2788 | delete tzenum; | |
2789 | } | |
2790 | ||
2791 | ||
51004dcb A |
2792 | UnicodeString& |
2793 | TimeZoneFormat::parseZoneID(const UnicodeString& text, ParsePosition& pos, UnicodeString& tzID) const { | |
2794 | UErrorCode status = U_ZERO_ERROR; | |
57a6839d | 2795 | umtx_initOnce(gZoneIdTrieInitOnce, &initZoneIdTrie, status); |
51004dcb A |
2796 | |
2797 | int32_t start = pos.getIndex(); | |
2798 | int32_t len = 0; | |
2799 | tzID.setToBogus(); | |
2800 | ||
57a6839d | 2801 | if (U_SUCCESS(status)) { |
51004dcb A |
2802 | LocalPointer<ZoneIdMatchHandler> handler(new ZoneIdMatchHandler()); |
2803 | gZoneIdTrie->search(text, start, handler.getAlias(), status); | |
2804 | len = handler->getMatchLen(); | |
2805 | if (len > 0) { | |
2806 | tzID.setTo(handler->getID(), -1); | |
2807 | } | |
2808 | } | |
2809 | ||
2810 | if (len > 0) { | |
2811 | pos.setIndex(start + len); | |
2812 | } else { | |
2813 | pos.setErrorIndex(start); | |
2814 | } | |
2815 | ||
2816 | return tzID; | |
2817 | } | |
2818 | ||
57a6839d A |
2819 | static void U_CALLCONV initShortZoneIdTrie(UErrorCode &status) { |
2820 | U_ASSERT(gShortZoneIdTrie == NULL); | |
2821 | ucln_i18n_registerCleanup(UCLN_I18N_TIMEZONEFORMAT, tzfmt_cleanup); | |
2822 | StringEnumeration *tzenum = TimeZone::createTimeZoneIDEnumeration(UCAL_ZONE_TYPE_CANONICAL, NULL, NULL, status); | |
2823 | if (U_SUCCESS(status)) { | |
2824 | gShortZoneIdTrie = new TextTrieMap(TRUE, NULL); // No deleter, because values are pooled by ZoneMeta | |
2825 | if (gShortZoneIdTrie == NULL) { | |
2826 | status = U_MEMORY_ALLOCATION_ERROR; | |
2827 | } else { | |
2828 | const UnicodeString *id; | |
0f5d89e8 | 2829 | while ((id = tzenum->snext(status)) != NULL) { |
57a6839d A |
2830 | const UChar* uID = ZoneMeta::findTimeZoneID(*id); |
2831 | const UChar* shortID = ZoneMeta::getShortID(*id); | |
2832 | if (shortID && uID) { | |
2833 | gShortZoneIdTrie->put(shortID, const_cast<UChar *>(uID), status); | |
51004dcb | 2834 | } |
51004dcb A |
2835 | } |
2836 | } | |
51004dcb | 2837 | } |
57a6839d A |
2838 | delete tzenum; |
2839 | } | |
2840 | ||
2841 | ||
2842 | UnicodeString& | |
2843 | TimeZoneFormat::parseShortZoneID(const UnicodeString& text, ParsePosition& pos, UnicodeString& tzID) const { | |
2844 | UErrorCode status = U_ZERO_ERROR; | |
2845 | umtx_initOnce(gShortZoneIdTrieInitOnce, &initShortZoneIdTrie, status); | |
51004dcb A |
2846 | |
2847 | int32_t start = pos.getIndex(); | |
2848 | int32_t len = 0; | |
2849 | tzID.setToBogus(); | |
2850 | ||
57a6839d | 2851 | if (U_SUCCESS(status)) { |
51004dcb A |
2852 | LocalPointer<ZoneIdMatchHandler> handler(new ZoneIdMatchHandler()); |
2853 | gShortZoneIdTrie->search(text, start, handler.getAlias(), status); | |
2854 | len = handler->getMatchLen(); | |
2855 | if (len > 0) { | |
2856 | tzID.setTo(handler->getID(), -1); | |
2857 | } | |
2858 | } | |
2859 | ||
2860 | if (len > 0) { | |
2861 | pos.setIndex(start + len); | |
2862 | } else { | |
2863 | pos.setErrorIndex(start); | |
2864 | } | |
2865 | ||
2866 | return tzID; | |
2867 | } | |
2868 | ||
2869 | ||
2870 | UnicodeString& | |
2871 | TimeZoneFormat::parseExemplarLocation(const UnicodeString& text, ParsePosition& pos, UnicodeString& tzID) const { | |
2872 | int32_t startIdx = pos.getIndex(); | |
2873 | int32_t parsedPos = -1; | |
2874 | tzID.setToBogus(); | |
2875 | ||
2876 | UErrorCode status = U_ZERO_ERROR; | |
2877 | LocalPointer<TimeZoneNames::MatchInfoCollection> exemplarMatches(fTimeZoneNames->find(text, startIdx, UTZNM_EXEMPLAR_LOCATION, status)); | |
2878 | if (U_FAILURE(status)) { | |
2879 | pos.setErrorIndex(startIdx); | |
2880 | return tzID; | |
2881 | } | |
2882 | int32_t matchIdx = -1; | |
2883 | if (!exemplarMatches.isNull()) { | |
2884 | for (int32_t i = 0; i < exemplarMatches->size(); i++) { | |
2885 | if (startIdx + exemplarMatches->getMatchLengthAt(i) > parsedPos) { | |
2886 | matchIdx = i; | |
2887 | parsedPos = startIdx + exemplarMatches->getMatchLengthAt(i); | |
2888 | } | |
2889 | } | |
2890 | if (parsedPos > 0) { | |
2891 | pos.setIndex(parsedPos); | |
2892 | getTimeZoneID(exemplarMatches.getAlias(), matchIdx, tzID); | |
2893 | } | |
2894 | } | |
2895 | ||
2896 | if (tzID.length() == 0) { | |
2897 | pos.setErrorIndex(startIdx); | |
2898 | } | |
2899 | ||
2900 | return tzID; | |
2901 | } | |
2902 | ||
4388f060 A |
2903 | U_NAMESPACE_END |
2904 | ||
2905 | #endif |